{"id": "train_45000", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "repo_scale_reasoning", "governance", "evaluation_metrics"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a8a3b6a08f942e90ce72a10a00b5232624efea48"}
{"id": "train_45001", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "documentation", "security_gates", "reproducibility"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ae8026f85da246e44ec83d458abb6e4f9b0c2821"}
{"id": "train_45002", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "tooling", "documentation"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "64f89c36267e78e28f6a566f41e30fb99a7df3ac"}
{"id": "train_45003", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "evaluation_metrics", "tests_are_truth", "documentation"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c15fe80f7757d997c7462a5237b1e97386dd19f3"}
{"id": "train_45004", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "repo_scale_reasoning", "ci_integration", "auditability"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "82b6e0a7e0f2f09d1b90d9d4e9c35e5c35005a1b"}
{"id": "train_45005", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "tests_are_truth", "security_gates", "ci_integration"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4049f651a0412d09c1f1ecf7a07247968f90479d"}
{"id": "train_45006", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "tooling", "tests_are_truth", "reproducibility"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ccb15be6b94b034e035e490eaf183e01419ea7cc"}
{"id": "train_45007", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "tooling", "evaluation_metrics", "security_gates"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7216f9a197dbe5f35347ab7e11b99330bb083ca7"}
{"id": "train_45008", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "tooling", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "647d333ad43a67a69ee3f0c43482bf4a7d75a56a"}
{"id": "train_45009", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "tests_are_truth", "ci_integration", "auditability"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0f64e609e26217eec6ce80212278f5e440336cb6"}
{"id": "train_45010", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "tests_are_truth", "auditability"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "85b2a3da3f6f0e0efd3ba7b5ae245c0b0d6537ed"}
{"id": "train_45011", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "cost_latency_tradeoffs", "auditability", "tooling"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b472e7ba8c596e5b4e5385b07cc47ed2a528ab7e"}
{"id": "train_45012", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "cost_latency_tradeoffs", "evaluation_metrics", "governance"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c833510da7d592e9ccbb3c3a44e438edbc67ff02"}
{"id": "train_45013", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "documentation", "ci_integration", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bfca0b8842cb159712e705b2a1029b38539af751"}
{"id": "train_45014", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "repo_scale_reasoning", "tests_are_truth", "security_gates"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f39952b259cc06aa5dab2d59b64c489efba089c9"}
{"id": "train_45015", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "cost_latency_tradeoffs", "tests_are_truth", "ci_integration"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d9999877e4864a8efd3330330f812c63e0d65081"}
{"id": "train_45016", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "deaaa043fcd1af2724abbeeaf1865918f59b7f4d"}
{"id": "train_45017", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "tooling", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "26bb7349ab6e0f3b7872cad5869d1565c14454d7"}
{"id": "train_45018", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "governance", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dfff49e23efdbb601e6d5146b152ad05b989d84a"}
{"id": "train_45019", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "security_gates", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "48d6e0d5ebf7a2413853b5176d5c5c2b03d0e796"}
{"id": "train_45020", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "evaluation_metrics", "tooling"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d6a5e2e9fbc16318e96f561de4775083d089f20d"}
{"id": "train_45021", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "governance", "documentation", "auditability"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fcc623e3139458c127b99c2dd4158252e6585d6b"}
{"id": "train_45022", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "governance", "documentation", "security_gates"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "61214afafc27331b46da9cbd0480fe9ec9e8d439"}
{"id": "train_45023", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "documentation", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "927269fbe6bd26d4f6e12fcc9834e7d866084e03"}
{"id": "train_45024", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "evaluation_metrics", "repo_scale_reasoning", "governance"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d38fc77de4f8de0dcd600a3afae2901bb9e3c068"}
{"id": "train_45025", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "governance", "evaluation_metrics", "reproducibility"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "03edab330632860a9b8a9a5ba4cc951691fa9ff0"}
{"id": "train_45026", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "ci_integration", "documentation", "reproducibility"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5c54f853b4ae41142178e3d179f85ebbdd4a1241"}
{"id": "train_45027", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "reproducibility", "documentation"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ec77247b4140e0a358bba62f026bf1e563417f64"}
{"id": "train_45028", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "security_gates", "ci_integration", "documentation"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "16f437f80782ca57bd5f0e768f08ba3894e60626"}
{"id": "train_45029", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "documentation", "tooling", "governance"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "84d089bbb457346de9b4d8ec4eba1183339ca3ad"}
{"id": "train_45030", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "repo_scale_reasoning", "tooling", "tests_are_truth"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4eab710dc6a37b3da2f50e58fdd7b43305b0adda"}
{"id": "train_45031", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "governance", "repo_scale_reasoning", "tooling"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9978bb0e05d6a2eca91ae94ba5bd1e24329a1db4"}
{"id": "train_45032", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "cost_latency_tradeoffs", "security_gates", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "552c541068fb117f30dd2805f9c92344cde22bb8"}
{"id": "train_45033", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "auditability", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9338258b568e59252aa092a02220ec72c5642476"}
{"id": "train_45034", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "ci_integration", "auditability", "tooling"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9b111f5707d3553afe68d04583d122f01f738da1"}
{"id": "train_45035", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "repo_scale_reasoning", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f4d1bfcfacdb48d8c12922d028c6f33d61c48564"}
{"id": "train_45036", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "governance", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3780dc897a9fe4ce30ff1ae642254492d04fa545"}
{"id": "train_45037", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "documentation", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7206e8255ed0eef9d5fb437b3b4dc055660f97d0"}
{"id": "train_45038", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "evaluation_metrics", "documentation", "reproducibility"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7c4ac632889760cda6649d35d3d05e7bc2bea8ce"}
{"id": "train_45039", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "tests_are_truth", "ci_integration", "auditability"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a1b5a608575a826aca1e683e534d10042c279d96"}
{"id": "train_45040", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "repo_scale_reasoning", "governance", "reproducibility"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4f7b90c0356a2b22a53b78f0d6cb48f314faf9d2"}
{"id": "train_45041", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "auditability", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d11301a105adc727e13935aa8ff11a856c0b63f5"}
{"id": "train_45042", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "reproducibility", "documentation", "tooling"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "018bad02fad1626c5e4607c45cf00c84d690e28e"}
{"id": "train_45043", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "ci_integration", "reproducibility", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "700536162c451fac61b628077ff8237d6a4179b8"}
{"id": "train_45044", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "documentation", "repo_scale_reasoning", "tooling"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f055e7be1132461e848ca088be52e8ab64c9e99e"}
{"id": "train_45045", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "tests_are_truth", "reproducibility", "tooling"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "82725934747f73149732924894f0e2aebf232b45"}
{"id": "train_45046", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "auditability", "tests_are_truth", "ci_integration"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f780144cfc67154a6d95f59b7d6b2cca53b75678"}
{"id": "train_45047", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "reproducibility", "tests_are_truth"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "802f198054e470f648ac04bd9d8fc12d398d09bb"}
{"id": "train_45048", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "security_gates", "governance", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc789a1afc67b390f4eb323429afeb01a6b33862"}
{"id": "train_45049", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "cost_latency_tradeoffs", "documentation", "security_gates"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9330186d3aa86af46b95c560567a63b531c7625d"}
{"id": "train_45050", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "tests_are_truth", "governance", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "34626fa2bb670662f92380e852e909a0a264df78"}
{"id": "train_45051", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "reproducibility", "governance", "auditability"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ba11092e960db27530c14ae3768dc5c988662989"}
{"id": "train_45052", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "evaluation_metrics", "governance"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2bc1e9a5f1dc11637166c0badf50539d23201d93"}
{"id": "train_45053", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "governance", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6f87fc9bd9cc6e4afec02ab50bd41a2832911c1a"}
{"id": "train_45054", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "ci_integration", "documentation", "governance"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5f66b211deba832bf0dcfe077df938e1c98f1fbc"}
{"id": "train_45055", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "auditability", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1f46b575a6c56749f1f80c0711675216eb16f50c"}
{"id": "train_45056", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "ci_integration", "security_gates", "reproducibility"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3a02daf2457754e0616b09ebc329715fc5ec6026"}
{"id": "train_45057", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "repo_scale_reasoning", "auditability"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "57741dcf1d88d5fca84d199d803f90635d82ff8f"}
{"id": "train_45058", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "tests_are_truth", "tooling", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8d0f46a1aa999978ee95c131731359e227e4af75"}
{"id": "train_45059", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "ci_integration", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e00dec58d15ad905ffa1c3c7ec11716ee079691f"}
{"id": "train_45060", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "repo_scale_reasoning", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "48d34b9bea2d4916906111d5dfb28ad6806481eb"}
{"id": "train_45061", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "tooling", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fa7956a2b994c1578614689bf44b37728d5eef35"}
{"id": "train_45062", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "tests_are_truth", "reproducibility", "documentation"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "deaa760f0c630dd40d5b53a955db82d1b6f27139"}
{"id": "train_45063", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "auditability", "repo_scale_reasoning", "governance"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b31ca137de3f070013a786fd059bdd8a9fea450c"}
{"id": "train_45064", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "tooling", "tests_are_truth", "security_gates"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b47f260bc969da5a85dfefff464e4bc35102af12"}
{"id": "train_45065", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "documentation", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a5fdbed797bc051134daaf3254bcef9e137a2731"}
{"id": "train_45066", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "repo_scale_reasoning", "documentation", "tests_are_truth"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ccdc9fd397e9189af5524c891ad92a12b04e59dc"}
{"id": "train_45067", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "auditability", "evaluation_metrics", "ci_integration"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4f006031a1aacc3c59a6e6e517a9ee47ec5264da"}
{"id": "train_45068", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "tooling", "repo_scale_reasoning", "auditability"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8bfbd1941e47de9a3cf87be24a7f3154c4bf02df"}
{"id": "train_45069", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "ci_integration", "reproducibility", "security_gates"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f5177f29620bb55c44c6d602fb2e0b7672a6e866"}
{"id": "train_45070", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "documentation", "tooling", "evaluation_metrics"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "57779c729ffa73c8f8bb18464ebce814f68c594d"}
{"id": "train_45071", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d2b3c1fb17628ea819a697d66f9294a7d06680e1"}
{"id": "train_45072", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "auditability", "security_gates", "documentation"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "af0c58b3aaf6c9b3b5f213c7489be54a9eec597f"}
{"id": "train_45073", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "auditability", "reproducibility", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "58c384731f7f15cbd15c03c17a612293ded0755f"}
{"id": "train_45074", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "tests_are_truth", "ci_integration"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fcd5efd991cc902bc4f1cd5710762fe61b05d191"}
{"id": "train_45075", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "security_gates", "tooling", "governance"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5fd55962ba6db5fb769d4369da27cfd2687de97d"}
{"id": "train_45076", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "evaluation_metrics", "tooling", "governance"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b013bbb18c29d13436738b7720fd53bdca30b2c0"}
{"id": "train_45077", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "governance", "reproducibility"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e7e37f1a020f3539e2759932f9c6507fb4c57ad8"}
{"id": "train_45078", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "reproducibility", "documentation"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5ff740fd2acbfa4253e185c860fea3a7955637c8"}
{"id": "train_45079", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "documentation", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f038357aca759a0749536718fdf5e3f6b51019df"}
{"id": "train_45080", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "evaluation_metrics", "tooling", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "70a7cd6b8a42353aadf9d89ca753c88aa61e5e51"}
{"id": "train_45081", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "auditability", "security_gates", "documentation"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "43e7f7fd41f5e6fe01d1c6d272419bf73c58a184"}
{"id": "train_45082", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "auditability", "security_gates", "ci_integration"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f99659e763e465e95cd4ebed0b1078bd815e62ac"}
{"id": "train_45083", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "security_gates", "documentation", "ci_integration"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0488a109981a2be35ad5284c9580713a65563b3a"}
{"id": "train_45084", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "repo_scale_reasoning", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aa53d874a1e352c096c217475778a35b663b4c6a"}
{"id": "train_45085", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "tooling", "documentation", "reproducibility"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e3c2a560225f78082d1e780fd45db2efd92306c4"}
{"id": "train_45086", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "auditability", "governance"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f7afb989a1984b75a994ea2617219394152adb19"}
{"id": "train_45087", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "documentation", "ci_integration", "reproducibility"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9b8727937fc2f7abb8533d971ba41ae703dd20f8"}
{"id": "train_45088", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "tests_are_truth", "security_gates", "governance"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0b5f8176ba09955e5147ef231860b99f6373294b"}
{"id": "train_45089", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "auditability", "security_gates", "ci_integration"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7aec61485d0f8d8ff638b5d56f88ae6666a75c02"}
{"id": "train_45090", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "auditability", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e9a0596065b50c8527bc95ef552b9335683485e4"}
{"id": "train_45091", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "ci_integration", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eb523a8b29665a687a541df0abd2fed1f6c78091"}
{"id": "train_45092", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["governance", "cost_latency_tradeoffs", "reproducibility", "auditability"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4510cefbcded174946a44533b33e21635eb275d3"}
{"id": "train_45093", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "security_gates", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "330abc29e723d1502b5ddbc9f1155f07a42e600e"}
{"id": "train_45094", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "evaluation_metrics", "ci_integration", "auditability"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ed5e7a7c5516d1d91589cca4332199150dfb7abe"}
{"id": "train_45095", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "repo_scale_reasoning", "auditability", "governance"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "59ee8b87d26e6796072404a0625d87a0aae38c7c"}
{"id": "train_45096", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "tooling", "ci_integration", "evaluation_metrics"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a7fb2b2c245c6106485f86af8c6aecf05a9d2016"}
{"id": "train_45097", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "tests_are_truth", "ci_integration", "documentation"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3218ee57c5faf33dda30d08b0617088ebf55d146"}
{"id": "train_45098", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "security_gates", "governance"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c4293ecfccd76480f0a0d53170bd73da9a603b31"}
{"id": "train_45099", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "tests_are_truth", "governance"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "522f5bf50d1566a6757bd43a684e0cf8d31f1b39"}
{"id": "train_45100", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "tooling", "security_gates"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8d2e549a485d5f636f61d40754772384eed9b0cd"}
{"id": "train_45101", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "tests_are_truth", "governance", "tooling"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fb91e9cc17b75d690bdc40473808ef118719e355"}
{"id": "train_45102", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "governance", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e5e845325cdb4f1650c50e213790630a2b8ca6c5"}
{"id": "train_45103", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "auditability", "tooling"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0c5754dde2990d0731ebca4402236b444ca0aaa7"}
{"id": "train_45104", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "ci_integration", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "16e1d0d5f08957d093d29e6d03fd13659dce77ec"}
{"id": "train_45105", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "777cec47d8aecdba531b235e17b25f0580d792cd"}
{"id": "train_45106", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "governance", "tooling", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "54a687b98b2e83e457d5abe9b4bc7dce5d4203d6"}
{"id": "train_45107", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "tooling", "security_gates", "tests_are_truth"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3ee3306d68d1b46a20b62d9b9b791b6a25c1f832"}
{"id": "train_45108", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "tooling", "auditability", "reproducibility"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9687604bafe64fa885857b0e8253766b43d76b42"}
{"id": "train_45109", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "tests_are_truth", "tooling", "reproducibility"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "25ef4592b9588f687e3530fcbecaf0857f7988fe"}
{"id": "train_45110", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "auditability", "documentation"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5d60681b51348ef287729949218cdd60125f6c17"}
{"id": "train_45111", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "auditability", "documentation", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "44d0468929bd3e7e3297b9942b6c0766d4384536"}
{"id": "train_45112", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "ci_integration", "governance"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "364f482807fc8a1b571ddc3e6a28ead00d4fd013"}
{"id": "train_45113", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "tests_are_truth", "documentation", "tooling"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ce580ea147a4bbc5231d21b24e4b7495dc059b6c"}
{"id": "train_45114", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "auditability", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ad423ad68f8aeb9024f9b29cca764fc6689d7e25"}
{"id": "train_45115", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "reproducibility", "security_gates", "ci_integration"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "91b1c2ab8b8b744f326a8dbe4c35db3214f8b2d1"}
{"id": "train_45116", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "tooling", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8e26c884f4b4436c7bc38d2422fbe2c2ebda9641"}
{"id": "train_45117", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "documentation", "ci_integration", "governance"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "048bf7a2f0903bb7d5f12be4d5196df390a06d0d"}
{"id": "train_45118", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "reproducibility", "evaluation_metrics", "documentation"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "06b8fec492a497a4309fb9d02a8a4d9e193e65e6"}
{"id": "train_45119", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "governance", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9f9e6c557ffd1b7fe4db5deb19f7ddfc6c882705"}
{"id": "train_45120", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "documentation", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "133ee636bb85b1697cca80272c3a367a66fb0d3e"}
{"id": "train_45121", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "documentation", "governance", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0e58caa79877ab53fb26d471acfa3750c4d9360e"}
{"id": "train_45122", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "ci_integration", "evaluation_metrics", "tooling"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c1f8f29917d2613975822d8a36c7aedf98de494c"}
{"id": "train_45123", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "documentation", "repo_scale_reasoning", "auditability"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1f15ab0e8f4db859209278ac672d2fafa985afde"}
{"id": "train_45124", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f04e0f6205d42a2dac1103baf7451939fbd45877"}
{"id": "train_45125", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "ci_integration", "auditability"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cf0991058eef10b482cae76b7ebca2bddeb7d69c"}
{"id": "train_45126", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "documentation", "governance", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fe40b4e5f2e268ace1aa388f0ec6bf331fdc6c25"}
{"id": "train_45127", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "tooling", "documentation", "tests_are_truth"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5abaa141de42746539be727019d41ec493bf591f"}
{"id": "train_45128", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "ci_integration", "governance", "auditability"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "600fd9e7554332a79b655f71c51f232bdb39d8ec"}
{"id": "train_45129", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "auditability", "ci_integration", "reproducibility"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aa7f4e6cef1e5bdcffb87e7f42760b06d17cf5fa"}
{"id": "train_45130", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "tooling", "governance", "reproducibility"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "afb3e4987f840517385ea0a1ced148400128c0fb"}
{"id": "train_45131", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "cost_latency_tradeoffs", "reproducibility", "ci_integration"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ad49338b7f8ef8f42b778c64f44f3cf9ae62296d"}
{"id": "train_45132", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "repo_scale_reasoning", "security_gates"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "077762be642420f7683c458f68bc884d3ce084e1"}
{"id": "train_45133", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "ci_integration", "auditability", "documentation"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b658a01f35ab93c6756adda5bcfb5ae2bbf1f34f"}
{"id": "train_45134", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "cost_latency_tradeoffs", "repo_scale_reasoning", "auditability"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "61ac2ad231f148ac878d9ecf6811907d19410e55"}
{"id": "train_45135", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "documentation", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c5df16c0c3f743e8c80aa96865abfa119772d872"}
{"id": "train_45136", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "ci_integration", "documentation"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e5914219d38a77dc96c9cb0b98c9df8a30829fae"}
{"id": "train_45137", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "tooling", "evaluation_metrics", "security_gates"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "923a4ef20390bdc995ec4af5005dcf6353864419"}
{"id": "train_45138", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "governance", "documentation", "security_gates"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "884de04cbe6cd9adff950e0cd38afcd5cfa24157"}
{"id": "train_45139", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "tests_are_truth", "repo_scale_reasoning", "security_gates"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "229ce157193cfd82efe56f041f4f43959db010d0"}
{"id": "train_45140", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "reproducibility", "security_gates"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bfd6ca8197c22169db04d67af567abe7d2169f40"}
{"id": "train_45141", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "ci_integration", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "42b0f395687d54b1d49afcffa8f6e3583c61bac5"}
{"id": "train_45142", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "tooling", "governance", "documentation"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "817bb89b0f21db34a0ce45aa84ef59b48b2f6c21"}
{"id": "train_45143", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "evaluation_metrics", "auditability"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "118ff4cadc75e1aa501b5f52dec203cdcf512f08"}
{"id": "train_45144", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "documentation", "governance", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4f5a7b4a4ad69f80b7e8b2e0c61d50dce2ca06fa"}
{"id": "train_45145", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "documentation", "evaluation_metrics"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7069487b6ae48630f03f919a0b1ac1905a21d12c"}
{"id": "train_45146", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "governance", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "29aa54dfafe4d0da87f525119d3fa0954f8ea5b6"}
{"id": "train_45147", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "governance", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2387b0448cc236189b71a208bd426721788553da"}
{"id": "train_45148", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "tooling", "cost_latency_tradeoffs", "governance"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "604cefb05179ffb3f4d8e7cabfcd9cbf324ff6d8"}
{"id": "train_45149", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "repo_scale_reasoning", "security_gates", "documentation"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "af7ef924754d390d7d81e6b2deb280e224347b3c"}
{"id": "train_45150", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "auditability", "tooling"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9e2c3b5240c0ec9727b42508ab715f23f23a136c"}
{"id": "train_45151", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "security_gates", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "98fc2fef1e6075c48b7e6a0708797dbbc668ab0c"}
{"id": "train_45152", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "documentation", "governance", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a00ca5237cf949fba83bbd9e3dac627d204656d5"}
{"id": "train_45153", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "tests_are_truth", "governance", "evaluation_metrics"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0979455f08781096c180d5cb055105c9922aecfb"}
{"id": "train_45154", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "tooling", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b360361ef83ab5ddefdbefa10b8e83892d0f42a5"}
{"id": "train_45155", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "security_gates", "reproducibility", "tooling"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ad22dbbd04fdbfdeb6ac521cb9c3a35075c81a7d"}
{"id": "train_45156", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "evaluation_metrics", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5ea155024e26b1c85c7b4e35bd7459575c2735b7"}
{"id": "train_45157", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["governance", "tests_are_truth", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "36fec1eb406b8fb922e3203be013609880edb6bb"}
{"id": "train_45158", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "tooling", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bb411324b0d265e5780e540c79c8ff41f0cc6ec9"}
{"id": "train_45159", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "tooling", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5bbeead378deeb68bc7282336fff0d0b011907fe"}
{"id": "train_45160", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "documentation", "tests_are_truth"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "99966bb51a7fc52504750ab8580b9d41981d5d65"}
{"id": "train_45161", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "tests_are_truth", "security_gates", "auditability"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ccf6723a5a2f5edc030547dd09d8aa732928623a"}
{"id": "train_45162", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "repo_scale_reasoning", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "69ec6251d327eac16e427befb43542a50d30ef4e"}
{"id": "train_45163", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "auditability", "ci_integration", "tooling"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "633c8fed64b1d09e5703d5bd7c27d7860d037c65"}
{"id": "train_45164", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "documentation", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "306afdd6977e3724e79e72d13f7527cd8ca19f8c"}
{"id": "train_45165", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "documentation", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0708930087b4e77ae2439ad3cc469b4a227b8e09"}
{"id": "train_45166", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "tests_are_truth", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0aa2072913acb43b1f1888ee5eecefb54851414f"}
{"id": "train_45167", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "cost_latency_tradeoffs", "reproducibility", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d1371523d0e3607b7a873aa3b3ca60c3187ea7e2"}
{"id": "train_45168", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "governance", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9fd5e5dddde41bcd8beba7fc4ca0796cab926ee5"}
{"id": "train_45169", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "tests_are_truth", "ci_integration", "reproducibility"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a3897b8f1b65858468953b8b7715ffc77b9b348d"}
{"id": "train_45170", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "evaluation_metrics", "ci_integration", "auditability"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a9184371a7bd185286025c0c8fe1ed4f4ed9beba"}
{"id": "train_45171", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "repo_scale_reasoning", "ci_integration", "evaluation_metrics"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "499a9c207db59c5d920a6954b543967e3d6531f3"}
{"id": "train_45172", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "ci_integration", "tooling", "tests_are_truth"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eef4c94d3ee8561c6c19543f7d43a960434323f6"}
{"id": "train_45173", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "documentation", "security_gates", "evaluation_metrics"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "001c35cf9757b9f6025b0d40d28ddeed1f04335e"}
{"id": "train_45174", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "tests_are_truth", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "adf6018b2a441bf89eeb45cc0090ee8c04f01316"}
{"id": "train_45175", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "repo_scale_reasoning", "security_gates", "ci_integration"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f84e0ddb44224321361cf802fc7f25980d1e3363"}
{"id": "train_45176", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "documentation", "repo_scale_reasoning", "security_gates"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1d2b8f0b43b7ce62ca4874c331492866b8637ed2"}
{"id": "train_45177", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "documentation", "governance", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e515026b7a82f2791988064ceeea5a79b6471054"}
{"id": "train_45178", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "tests_are_truth", "cost_latency_tradeoffs", "governance"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4e8c485f1168a3fb24b0696c7087d6ec75e9daaa"}
{"id": "train_45179", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "ci_integration", "repo_scale_reasoning", "auditability"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4997f749ddd3e1df6842f9bd5bd62c8860613d19"}
{"id": "train_45180", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "tooling", "security_gates", "tests_are_truth"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ed102da6b980a704e63a78db8bdc7ef66e3addce"}
{"id": "train_45181", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "security_gates", "tooling", "reproducibility"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "927c897e967378eb884fa3176182387a0db84d57"}
{"id": "train_45182", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "reproducibility", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "992a1ce2e0795477ec633c5b67fedb2b4c3767a1"}
{"id": "train_45183", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "cost_latency_tradeoffs", "repo_scale_reasoning", "auditability"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a1b5a608575a826aca1e683e534d10042c279d96"}
{"id": "train_45184", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "documentation", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1d0ee659d3afdacd8926114c3c56ef0c800e17e1"}
{"id": "train_45185", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "tests_are_truth", "evaluation_metrics", "ci_integration"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ec2b4a301d8585a9e37ad083f84e2260e43c6933"}
{"id": "train_45186", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "tooling", "security_gates", "auditability"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ef2327849b30a0c40e50fb2900cc7555338f69bf"}
{"id": "train_45187", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "tooling", "ci_integration", "tests_are_truth"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d5002a77096709517dd028501473baa241239038"}
{"id": "train_45188", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "governance", "ci_integration", "evaluation_metrics"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c3ae75400d50d4860bdf69769bbb28550c4b774d"}
{"id": "train_45189", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "repo_scale_reasoning", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a8f5c4fe5fa9aeb26505ea433c52eff3d00d0d23"}
{"id": "train_45190", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "tests_are_truth", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5b16b1206e2ff34af52d76cc1796f90968e74f42"}
{"id": "train_45191", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "security_gates", "reproducibility"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fe059374962e7d7762d8b245158f3ba416b3cb4d"}
{"id": "train_45192", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0698fa67edd5751fbf83251bd1b07aae790c155a"}
{"id": "train_45193", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "governance", "ci_integration", "reproducibility"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6e49735f10d3ba1809dc1a4bb02915bdd896a911"}
{"id": "train_45194", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "documentation", "tooling", "reproducibility"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "26ead9db86f3aba7b894c44f8e4881d35db891d0"}
{"id": "train_45195", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "evaluation_metrics", "repo_scale_reasoning", "documentation"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "687e8d8d01a44dc7dfac427ec4509e9fdd35a821"}
{"id": "train_45196", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "repo_scale_reasoning", "auditability", "documentation"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7adcef4f1063e144319d58a227244eb57d0cb9dd"}
{"id": "train_45197", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "evaluation_metrics", "security_gates", "auditability"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d614ba08a0f36713fd13e7fb3b5820e7a854f80a"}
{"id": "train_45198", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "documentation", "auditability", "security_gates"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "796c073f6afcb8b17326181a8408b3d5c8006507"}
{"id": "train_45199", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "documentation", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "479d3cbe61db2c1b42f6efbe2507c74849476e23"}
{"id": "train_45200", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "documentation", "tests_are_truth", "security_gates"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e2378f28ca40770f03ccb3a26729e39b0330816d"}
{"id": "train_45201", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "auditability", "ci_integration", "tests_are_truth"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8b46d1b47104461964089064951341edf22b5f41"}
{"id": "train_45202", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "governance", "documentation"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ed3fcbec41ae5e6597d97bbaed9ec1db4b9f0c35"}
{"id": "train_45203", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "security_gates", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9f924935010f376e2ca28ec2cb6fdfee788e2b18"}
{"id": "train_45204", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "documentation", "evaluation_metrics"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b6617ab6d29f1670f0dff61efa82c8d3bf10d600"}
{"id": "train_45205", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "tests_are_truth", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1bb3c2643c02a265ece8fdc7efcd050bd87d7af1"}
{"id": "train_45206", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "auditability", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0ad3c82b61f50932ddc34e36f96a055c053ad734"}
{"id": "train_45207", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "cost_latency_tradeoffs", "ci_integration", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "836316331875fca30fbc6416e36e1343dfef71dc"}
{"id": "train_45208", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "documentation", "evaluation_metrics", "ci_integration"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ac851743a9556ee8dd5bea6192e9e612ca03caa7"}
{"id": "train_45209", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "auditability", "documentation", "evaluation_metrics"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8f09b499fa2456b41b7696a21d520caba0d1e363"}
{"id": "train_45210", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "governance", "reproducibility", "evaluation_metrics"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b3d1a6a29d597e9567ade3c72d357854da9cc81c"}
{"id": "train_45211", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "auditability", "tooling", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d2b3c1fb17628ea819a697d66f9294a7d06680e1"}
{"id": "train_45212", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "documentation", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1a1944ce4e90b6102cb5298900ecd670022c9a34"}
{"id": "train_45213", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "auditability", "ci_integration", "tests_are_truth"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "17636408f0e979eb49bc4a53ac988130b3bf2e9c"}
{"id": "train_45214", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "security_gates", "repo_scale_reasoning", "governance"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d76fe65abb0a37c4b0d73bc7e6d2b598054c4138"}
{"id": "train_45215", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "evaluation_metrics", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "030437eff15793829f638aad99011b3af30164bd"}
{"id": "train_45216", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "auditability", "ci_integration"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "42d31911f3aaf24cef3af993ce2188ae3e901a5e"}
{"id": "train_45217", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "security_gates", "auditability", "tooling"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0201feb751ef8426db71c8b00cce6e688d33ddbf"}
{"id": "train_45218", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "repo_scale_reasoning", "auditability", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "605accf2454b258c641bd759e706789828dfb999"}
{"id": "train_45219", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "reproducibility", "documentation"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2b81852f8502544d1972cb859357af250048d826"}
{"id": "train_45220", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "auditability", "security_gates", "governance"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "26d89b049c843d1cd0504067a60c2345706ee631"}
{"id": "train_45221", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "documentation", "security_gates", "evaluation_metrics"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6c5e7ba70394951afd8a1e86d824a520f4d2dc80"}
{"id": "train_45222", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "tooling", "reproducibility"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a805c85028bec6667371ac8864f597831a9043fe"}
{"id": "train_45223", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "documentation", "reproducibility", "tooling"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "df8c22acfd3ff5b991156c3360857f1e191d2314"}
{"id": "train_45224", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["governance", "ci_integration", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "76becc765b4c2d3b08d3d4f8bb51364a90e82ee4"}
{"id": "train_45225", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "repo_scale_reasoning", "reproducibility", "auditability"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8ab2ab39224209b30c804df3599ac42f2c962420"}
{"id": "train_45226", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "evaluation_metrics", "reproducibility", "security_gates"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d2e7b9087f559782fa99f720b4613b1d2b708896"}
{"id": "train_45227", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "documentation", "reproducibility", "security_gates"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "13a6cb0a7a8ba5e0d6cae9d2675a99d621978212"}
{"id": "train_45228", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "auditability", "security_gates"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "83651bd644a0d4faeb46f9741b8e0c9d53c274cb"}
{"id": "train_45229", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "tooling", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0fe53fc398818e88c72a86ca11180f438aef4e22"}
{"id": "train_45230", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "reproducibility", "documentation", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fe1db242540fd4ecb0494502653be21e718064ef"}
{"id": "train_45231", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "security_gates", "tooling", "auditability"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ebdf2da374b96f062706d6b58f048b0468ec5a95"}
{"id": "train_45232", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "documentation", "tooling", "ci_integration"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0a1238a2284adfafd106905f3a1e469042eed720"}
{"id": "train_45233", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "security_gates", "ci_integration", "governance"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7df8d565164079cc75d4baf34fb027d998147f78"}
{"id": "train_45234", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "documentation", "auditability"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e4984e5f52c70514014122122e51279b3de93448"}
{"id": "train_45235", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "tooling", "documentation"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c4a13ab09a9df5f45bf5289e58abdede4e5109ce"}
{"id": "train_45236", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "evaluation_metrics", "governance", "ci_integration"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9f77a2675034f120563321d705d443daa44bf3ff"}
{"id": "train_45237", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "governance", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3eb58e7a2706a1e2993bef092f05399255ade3a4"}
{"id": "train_45238", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "auditability", "tests_are_truth", "governance"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "82725934747f73149732924894f0e2aebf232b45"}
{"id": "train_45239", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "auditability", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ae882082964c5c65a7764440c6ec9c0858710325"}
{"id": "train_45240", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "tests_are_truth", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2000560a3532c59d9e88e3fd110df792183f5237"}
{"id": "train_45241", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "tests_are_truth", "governance", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9ed3ea60f58c07f6f70f7589b444c74225fb5eef"}
{"id": "train_45242", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "governance", "tests_are_truth", "auditability"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "31717dde7d7dbcdc0c2d2d6013c4fbfab9159d83"}
{"id": "train_45243", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "tests_are_truth", "reproducibility", "ci_integration"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0ee20d4ab15c0f0f77b3f9d924e969745891661d"}
{"id": "train_45244", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["governance", "tests_are_truth", "security_gates", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "db61633bd16b10722caf4833ea428d27853021c0"}
{"id": "train_45245", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["governance", "tooling", "auditability", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3f34430a52cf3d075e27e4640f33a6d94feb8987"}
{"id": "train_45246", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "evaluation_metrics", "repo_scale_reasoning", "documentation"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "49ac79c34a8cd0ca0abfe54a1146fe411d296de6"}
{"id": "train_45247", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "repo_scale_reasoning", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "03643bf99d584459163b43f4f2261f2da29f5d76"}
{"id": "train_45248", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "documentation", "tests_are_truth"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8414c469959bb5adaf723dc682ef0d0949297b36"}
{"id": "train_45249", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "repo_scale_reasoning", "evaluation_metrics", "security_gates"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "833c75eeeeabb1febf3bbe4e90162bf9da69570a"}
{"id": "train_45250", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "evaluation_metrics", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3d0c18bf40b34e4822921ceb7fae2ee5a220a520"}
{"id": "train_45251", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "governance", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4f4780fd40ac5145274cb2eb92be4c2a06ec6d7c"}
{"id": "train_45252", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "cost_latency_tradeoffs", "reproducibility", "documentation"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5d2ae2db6a75a67154dc3db0e8ab1c2ac4e99e70"}
{"id": "train_45253", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "reproducibility", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2444eb9074042830f72c8e15bafc69abedb68106"}
{"id": "train_45254", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "repo_scale_reasoning", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "67f0d354f3920b4ccd0bef61122f7fc5a01bc1dc"}
{"id": "train_45255", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "tests_are_truth", "security_gates", "documentation"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8838b9665a90f082a348223bcb12c688bb946153"}
{"id": "train_45256", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "documentation", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d8b4935d7fe1124ee0962bfd28ea2db0d4a43fd7"}
{"id": "train_45257", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "tests_are_truth", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e515026b7a82f2791988064ceeea5a79b6471054"}
{"id": "train_45258", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "security_gates", "reproducibility", "tests_are_truth"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "89f61fd0aa7c2288cbdc1e521aa5d8ab17c05080"}
{"id": "train_45259", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "tooling", "auditability"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4e712233f7ea7299114cae93731bbb436b9d8b5b"}
{"id": "train_45260", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "reproducibility", "documentation", "tests_are_truth"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cf8dd3ed38df1051f024698044605057517c4ae7"}
{"id": "train_45261", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "tooling", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d631179debd26101feec17990cde292ba7f461a7"}
{"id": "train_45262", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "security_gates", "documentation", "reproducibility"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "98ae2c4b9c0e98a88cd3ae8a58824bd9cfdc6cdd"}
{"id": "train_45263", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "security_gates", "governance", "ci_integration"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "87e291ad9394b7d1c91ce6895845f06deafb7c21"}
{"id": "train_45264", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "auditability", "documentation", "evaluation_metrics"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4fd8ab2ba2746ec890aba76af68a26000f0e810d"}
{"id": "train_45265", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "security_gates", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "26a35abf1ae588e3cd91cf96dd8bbec2ffb5fae7"}
{"id": "train_45266", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "repo_scale_reasoning", "auditability", "reproducibility"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "509637f12b6400de184d44275f6144bc7cf57a7f"}
{"id": "train_45267", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "auditability", "tests_are_truth", "tooling"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eb6aff26bd8c06e4f7d5698eb20299fc4ebaa1c7"}
{"id": "train_45268", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "ci_integration", "repo_scale_reasoning", "auditability"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "afaad93acbc3cb94b230f3922f43c6c58d4bf3aa"}
{"id": "train_45269", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "tests_are_truth", "reproducibility", "auditability"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c36bf0b2bae83f3b41b0cb5d1f7e15436e180964"}
{"id": "train_45270", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "documentation", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b4782e2c2391103d0ac31cdef9edd2a9addda4e5"}
{"id": "train_45271", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "reproducibility", "tests_are_truth", "governance"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "562517bf530912a8fdebe77ecf12feb3934e42bc"}
{"id": "train_45272", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "tests_are_truth", "governance", "auditability"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "efe9c136ab6a36ce5c3fb4bb2597c27327eb7974"}
{"id": "train_45273", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "documentation", "governance", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4bfbb96a30d7a7fd5836221eda7d52a4f959b4b3"}
{"id": "train_45274", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "ci_integration", "reproducibility", "auditability"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d0a27b9298748f87f9b797a9dfb4a2491b5cd80f"}
{"id": "train_45275", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "governance", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ec093006cc408c3a9e378778de2c8fe8cb213fb5"}
{"id": "train_45276", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "cost_latency_tradeoffs", "documentation", "repo_scale_reasoning"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9fd5e5dddde41bcd8beba7fc4ca0796cab926ee5"}
{"id": "train_45277", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "security_gates", "cost_latency_tradeoffs", "governance"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "67cd4b473d373b5ba5f4c46427b2b771f1342e7c"}
{"id": "train_45278", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "auditability", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f2ac60b6fe7de8e00d557d4ab273cb40f680d3be"}
{"id": "train_45279", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "auditability", "governance", "tests_are_truth"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9066ed9e134395ae20289263f3d65fe584dc031f"}
{"id": "train_45280", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "governance", "tests_are_truth"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1ce6641fe892c3f9b874c5e34151eaff4861b49a"}
{"id": "train_45281", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "evaluation_metrics", "tests_are_truth", "auditability"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fcafe368dcdaf134b995d8d3097f65b86e4d21fd"}
{"id": "train_45282", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "evaluation_metrics", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f4643c2174fe239473eebab5362cf2291c6af1d1"}
{"id": "train_45283", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "governance", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "12f14dbb7c5d5646e20e6c2e98095fe19c287ef3"}
{"id": "train_45284", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "reproducibility", "tests_are_truth", "security_gates"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a392ab75e1ebe42701e51f994d51b18a65ea816b"}
{"id": "train_45285", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "reproducibility", "repo_scale_reasoning", "security_gates"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cf184903034db28f83f705803bb440897e823c3c"}
{"id": "train_45286", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "security_gates", "tests_are_truth", "auditability"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "702fad50b2f2ce9cdf9c7b5fa47babe2b61c9af9"}
{"id": "train_45287", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "reproducibility", "repo_scale_reasoning", "tooling"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8d54370686b8e28272cfef18eec1524e81038b78"}
{"id": "train_45288", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "governance", "security_gates", "reproducibility"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "702fad50b2f2ce9cdf9c7b5fa47babe2b61c9af9"}
{"id": "train_45289", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "evaluation_metrics", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a568f36437004accdafea557d5ac63698cb836ea"}
{"id": "train_45290", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "repo_scale_reasoning", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "defff092d880df35613fa964cb45994a7cf43dc5"}
{"id": "train_45291", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "reproducibility", "governance", "auditability"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ff51666b6458a8b36ba937fd16fd5f1c3c4f535c"}
{"id": "train_45292", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "security_gates", "repo_scale_reasoning", "auditability"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0f426809ff0893d2d1c51aa4d609a761ee5f5f24"}
{"id": "train_45293", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "cost_latency_tradeoffs", "ci_integration", "security_gates"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "735ec54a1abbb6e1489bebaf9fe95f9ff5c9a9e4"}
{"id": "train_45294", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "ci_integration", "governance", "security_gates"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ba914c3b2dcb00a5b232282865bad9166307df4c"}
{"id": "train_45295", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "auditability", "tests_are_truth", "security_gates"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "36f4f864dfaa606955bc52885c7784ca154bfdb0"}
{"id": "train_45296", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "tooling", "security_gates", "tests_are_truth"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "09d2adc3f4a5307ed44df760b7de2c27985abc25"}
{"id": "train_45297", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "auditability", "governance", "security_gates"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e3953fdb56b5350349322b144701fe2c8c18af65"}
{"id": "train_45298", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "tooling", "documentation", "tests_are_truth"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d1c9c0c530434a1e3c13f07e699d09f379fde63b"}
{"id": "train_45299", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "auditability", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "03358510482d5fbbb6c6a7ee77d005ec73456fd4"}
{"id": "train_45300", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "tests_are_truth", "ci_integration", "documentation"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c109c997869b92d9f7ede5ba1d193955b9217eff"}
{"id": "train_45301", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "auditability", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e82a72528aba46072ea37aa5c17bb01c3e05f601"}
{"id": "train_45302", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "security_gates", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0f08ae9b4bf3a5eede7d292ee19c7ffd8478521b"}
{"id": "train_45303", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "auditability", "evaluation_metrics", "governance"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1b8db5cf28db4803186f3cb7ca43b01b8c9d5852"}
{"id": "train_45304", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "governance", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "50e25a64e109256307f21529b8c93581e636f603"}
{"id": "train_45305", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "documentation", "tooling", "governance"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6f311ce2d4b70cd3b4c664028609ddba57cb7698"}
{"id": "train_45306", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "auditability", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "37505396e98c4604f347329673a5692b974876bf"}
{"id": "train_45307", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "tooling", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2521ab1df325d1070025b73ab5d8e2d86a3346a7"}
{"id": "train_45308", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "tests_are_truth", "auditability", "security_gates"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "84c2ceb1f1d5cf8323885643db6c9e9d5d248933"}
{"id": "train_45309", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "security_gates", "governance", "ci_integration"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f3185eb9f5c784ec56ec5f663c5a0da177c4e6d5"}
{"id": "train_45310", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "repo_scale_reasoning", "documentation", "tests_are_truth"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "73adcc2948c04205660e3912bb786f1c42004000"}
{"id": "train_45311", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "security_gates", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e4ba1603754dae4763f9fd73758fae2ea1ae0d15"}
{"id": "train_45312", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["governance", "documentation", "security_gates", "auditability"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aebabc62b6627f210d0d4753bfa0c8f92fdaa418"}
{"id": "train_45313", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "reproducibility", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3d691551df51894e5d07829220e97b2338fe2aad"}
{"id": "train_45314", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "documentation", "evaluation_metrics", "ci_integration"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d33c8a4c431615e315c4376375870f4ca7276b8d"}
{"id": "train_45315", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "tooling", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "43a3f9528f0665cc38aedb2186fd249a807fb6e1"}
{"id": "train_45316", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "ci_integration", "tooling", "documentation"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d6eb05abb148bfc06ded41fb5926d45d9b96f39a"}
{"id": "train_45317", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "security_gates", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "de7990b9141bae75582d13fe01d826c674b79114"}
{"id": "train_45318", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "reproducibility", "auditability", "governance"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8fab15aa765e6e0cdf23418f595c88ddf715bb75"}
{"id": "train_45319", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "ci_integration", "governance", "documentation"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5891b30153211125ea811bb0460c83c6b62935ee"}
{"id": "train_45320", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "reproducibility", "ci_integration"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6eb5d51e37b75c0aa21559980e3186f8b6f00310"}
{"id": "train_45321", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "governance", "repo_scale_reasoning", "security_gates"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "931322d5c22cdbdb5db412d784b6317a942fec89"}
{"id": "train_45322", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "tooling", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a62705be509bab230b406e90a4cfa81ad2492749"}
{"id": "train_45323", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "tests_are_truth", "documentation", "ci_integration"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "329389663f7c042b2976539f75b65b06b78d36e9"}
{"id": "train_45324", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "governance", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "04103777a89c7dbbe151b73a45bc6afa09beff5d"}
{"id": "train_45325", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["governance", "documentation", "tooling", "reproducibility"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3f6d145acf8a931d4cca556d8ed13451903d80c5"}
{"id": "train_45326", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "governance", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "689a18fab7f7642b4194ef45f8e78db154409936"}
{"id": "train_45327", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "auditability", "documentation", "ci_integration"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2ce53f54ace1dfe9ce790194f32f92c8fd6eb664"}
{"id": "train_45328", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "documentation", "repo_scale_reasoning"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8ae83f84c63a2c152a5149fff8cf22830889b3e7"}
{"id": "train_45329", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "ci_integration", "documentation"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "382a394f7dfb88e9b3f330f20d968435cfce1e9d"}
{"id": "train_45330", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "repo_scale_reasoning", "documentation"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4ffb1266ca6719dade21e2c00aafd919c5f94774"}
{"id": "train_45331", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "security_gates", "governance"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9738f03f5fe704480e854344ffb1b0d087041feb"}
{"id": "train_45332", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "auditability", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b05efc2fdacbc5f1848a5c19906f952ac009f29d"}
{"id": "train_45333", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["governance", "tooling", "repo_scale_reasoning", "auditability"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "314b0b11efc11918408b5b32a96dceeda4aeb32a"}
{"id": "train_45334", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "repo_scale_reasoning", "tooling", "tests_are_truth"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "61d73761fc5f3ad47c90eb31a40a8c95cea0bc25"}
{"id": "train_45335", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "evaluation_metrics", "documentation", "tests_are_truth"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ba5d89071932df315ca511e08a237aa371f5e945"}
{"id": "train_45336", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "governance", "security_gates", "tooling"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d8be6fd34b4d0132f13ce921ec751c73d62878c6"}
{"id": "train_45337", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "tests_are_truth", "reproducibility", "documentation"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "885366d1dec0acf42177e53ac3dafc30f60d31a3"}
{"id": "train_45338", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "evaluation_metrics", "repo_scale_reasoning", "auditability"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7b5b04fbd52c3c26eee3b856ec21e885472ac72b"}
{"id": "train_45339", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "evaluation_metrics", "tooling", "ci_integration"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c948da4693d7b204e9a1d25dc3574bd78cf1ef27"}
{"id": "train_45340", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "evaluation_metrics", "repo_scale_reasoning", "governance"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c38bc87bfd757e761517a28b430665eeb36a4472"}
{"id": "train_45341", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["governance", "cost_latency_tradeoffs", "documentation", "auditability"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0fd3fda5e2caefc6cc4dcf9f062ef682a5725550"}
{"id": "train_45342", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "governance", "auditability", "reproducibility"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c6e65d8391f361c5556d501bbfe96dd42c763750"}
{"id": "train_45343", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "tooling", "ci_integration", "auditability"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0f2560a2f23b0e7cd7c6c423ad44048366d070cb"}
{"id": "train_45344", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "tests_are_truth", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2f310d4f4bf247196a93b5e4e8c1e7242145256a"}
{"id": "train_45345", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "ci_integration", "security_gates"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b49df7a7c521b798d3af2cc6a3a195e8493ff802"}
{"id": "train_45346", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "auditability", "tooling", "reproducibility"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3b4a0081d9f4a1dc9fb4d3e5b5699929a95a2529"}
{"id": "train_45347", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "auditability", "tests_are_truth", "documentation"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1919f9e45ecb2149fb5c357258fb396325753d3c"}
{"id": "train_45348", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "governance", "tooling", "documentation"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b4057df8db491dbef69429bd733efad19e3db0c6"}
{"id": "train_45349", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "evaluation_metrics", "tooling", "governance"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c2f00e56169d6a9c40c564f6f4672c831ae62aee"}
{"id": "train_45350", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "security_gates", "documentation"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6775ea14959ea36586a8c1e411d740f3c995661f"}
{"id": "train_45351", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "tooling", "repo_scale_reasoning", "security_gates"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ed9e5c4a332683b398a449b6ae543989dac54635"}
{"id": "train_45352", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "governance", "evaluation_metrics", "auditability"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6d258ded644a58397413b4022b02763b8f40d6e3"}
{"id": "train_45353", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "repo_scale_reasoning", "evaluation_metrics", "governance"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "69cc06b49c8c54a9f1ed35bfe38fb46e4b4b1129"}
{"id": "train_45354", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "governance", "ci_integration", "tests_are_truth"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c92dac616e663a45d891970641ca850743411798"}
{"id": "train_45355", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "ci_integration", "documentation"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7fb18aa89a4958ce92a72c31532176240fb38b4e"}
{"id": "train_45356", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "tests_are_truth", "ci_integration", "reproducibility"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e14e1ba7e96e93366c7fdc439f32215b8c0819d9"}
{"id": "train_45357", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "tooling", "evaluation_metrics", "security_gates"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dddfbf8833bac68634a67567a3761a64b6a39a1f"}
{"id": "train_45358", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "tooling", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "12a822a3c5c1bda031d786020584674697c3d8b7"}
{"id": "train_45359", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "ci_integration", "tooling", "repo_scale_reasoning"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "28d357837f996a261bec8edca00884f956573312"}
{"id": "train_45360", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "governance", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "91b1c2ab8b8b744f326a8dbe4c35db3214f8b2d1"}
{"id": "train_45361", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "tooling", "reproducibility", "ci_integration"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "561b1e0167b58c3407bcb9b2f9601f71a9cff423"}
{"id": "train_45362", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "governance", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "60ec3ce97133b4639e88410bec50a874d76a1d50"}
{"id": "train_45363", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "ci_integration", "evaluation_metrics"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9c783515726906efaed1e527664f9a010c30370f"}
{"id": "train_45364", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "evaluation_metrics", "tests_are_truth", "auditability"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1a38caf9a3c30fb4dffd0f2a714823c29aefc113"}
{"id": "train_45365", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "reproducibility", "documentation", "tests_are_truth"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "39c93383f5dd8807ddfb4f8f59d019eb2fac15e4"}
{"id": "train_45366", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "tests_are_truth", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d26a40ac8012fe2a51a2119c63f4c43752dcba7b"}
{"id": "train_45367", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "evaluation_metrics", "tests_are_truth", "documentation"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ebdca48e7871a75d461ab19198ebfe761d298328"}
{"id": "train_45368", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "reproducibility", "security_gates", "governance"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9a3e7d6b49e478641116d3ba407c05d40efb3df3"}
{"id": "train_45369", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "evaluation_metrics", "repo_scale_reasoning", "governance"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cda0093a46aa332ddf54bfd6d4699919d6b0cc41"}
{"id": "train_45370", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "cost_latency_tradeoffs", "governance", "security_gates"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ca93159e710d65756f7b1345109941be0629dad2"}
{"id": "train_45371", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "evaluation_metrics", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f0b713c6cd9de7891b27449a08bc552013ca71de"}
{"id": "train_45372", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "ci_integration", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "999978a49f09a172f524e24bea705595ea1e4177"}
{"id": "train_45373", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "cost_latency_tradeoffs", "tooling", "auditability"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7f1027d864ba6adc40b4b63aae28457cdef63384"}
{"id": "train_45374", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "documentation", "governance", "auditability"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "752c5e3c5ca0a44fc5013118bbf9ae8a84c8cb13"}
{"id": "train_45375", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "governance", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "636b7f1724d8cdc3d6bf49dcfab9ddbdad47abb5"}
{"id": "train_45376", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "ci_integration", "evaluation_metrics", "reproducibility"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d7259a77f24917652d42abf5f061df942d08a6d3"}
{"id": "train_45377", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "governance", "security_gates", "documentation"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "faea17072b31089a1d0d44245c85c11889e1c4e8"}
{"id": "train_45378", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "reproducibility", "governance", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ec7bbaa2750da6eb783bf4511ac59c2b7a8b3885"}
{"id": "train_45379", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "tooling", "evaluation_metrics", "reproducibility"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8098ccd97bad9fe8889d88ee0126b72a7f3b7601"}
{"id": "train_45380", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "repo_scale_reasoning", "security_gates", "reproducibility"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e761b08786917b7a5ba908bbae30c2b23b3ab12c"}
{"id": "train_45381", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "evaluation_metrics", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "900ce1500c06587a5753223f28e6ce11c0c55265"}
{"id": "train_45382", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "evaluation_metrics", "auditability", "documentation"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e1d50d89ab4b7b529c2cbf562b2517d585969033"}
{"id": "train_45383", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["governance", "evaluation_metrics", "tooling", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7a17e80f61edd6d265ebdb6727f6ad8a9d372f39"}
{"id": "train_45384", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "governance", "evaluation_metrics"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8b59d1bc448e936582d93e2a782a68cdbc4e3068"}
{"id": "train_45385", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "documentation", "ci_integration", "reproducibility"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "908e8b638a6ac874d6852b13a24cb8cdeeaaff7c"}
{"id": "train_45386", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "auditability", "reproducibility"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "685c27f81bf4670125e4bb87d5c92484b0ebb24f"}
{"id": "train_45387", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "reproducibility", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ff33eef93bda1de75c1b1bfc4cfe6920c1f58ee4"}
{"id": "train_45388", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "tooling", "governance", "security_gates"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b970e5694ac346f3af910b118a6b8a77aa4db3f0"}
{"id": "train_45389", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "evaluation_metrics", "repo_scale_reasoning", "auditability"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6a5330be73c8ad7c9effa3647b1e6da76d1e66b5"}
{"id": "train_45390", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "documentation", "auditability", "tests_are_truth"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1a1ad4babfc397e0bbfc0b6cdd96c1eb4966f39d"}
{"id": "train_45391", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "tooling", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a33166b6a2f246fc6ed666bc689150df2e6a2d07"}
{"id": "train_45392", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "reproducibility", "evaluation_metrics"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5d66c69c4508b3b737cd320b6b2e649b85214005"}
{"id": "train_45393", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "ci_integration", "governance"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6008c87e049d38b06952fefa27c204d806508eed"}
{"id": "train_45394", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "cost_latency_tradeoffs", "ci_integration", "reproducibility"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7857005340fb94de39f5ad336cf1cfe82d26ea37"}
{"id": "train_45395", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "cost_latency_tradeoffs", "tooling", "governance"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9dc5317574ed5cee21e721ebab6064b7c6a91036"}
{"id": "train_45396", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "governance", "ci_integration", "reproducibility"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d0275b71c53e24ad8425d520b4406cddec9a7ec9"}
{"id": "train_45397", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "repo_scale_reasoning", "security_gates", "governance"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5b2b5ed2d4fcb0933a14f0ceb1f55fea653ec0f1"}
{"id": "train_45398", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "tests_are_truth", "reproducibility", "evaluation_metrics"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4cbfe914385610b18328fc85b67ed98f5b9f2349"}
{"id": "train_45399", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "ci_integration", "reproducibility", "tooling"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3c84b885c8c76a20e522e7d4ed55dbda1371c563"}
{"id": "train_45400", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "ci_integration", "security_gates", "reproducibility"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6aec82bf6f2abd0ae99fb41d6db090f2a1a28f8b"}
{"id": "train_45401", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "tests_are_truth", "repo_scale_reasoning", "auditability"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "45eb02cd139f3f032e37df466f2a157044bd9e98"}
{"id": "train_45402", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "tests_are_truth", "repo_scale_reasoning", "tooling"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c7bc344e9c7111b0e39b7f9ed69a9ea5ac9c3f47"}
{"id": "train_45403", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "security_gates", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6048d55dd2325401c3731cce61c3808e9d680bc6"}
{"id": "train_45404", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "governance", "tests_are_truth", "reproducibility"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9851eb4b6ce909e82704cdfb39868f1f3e4cf184"}
{"id": "train_45405", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "governance", "reproducibility"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "851491f0b7253cc9457031cb756427de7fd88c12"}
{"id": "train_45406", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "repo_scale_reasoning", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bf8e87411264bceae0f33270cf1164c8c08b1982"}
{"id": "train_45407", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6abaa1df667758d6c0fde4b5a3aeff960e7fbb71"}
{"id": "train_45408", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "reproducibility", "tests_are_truth", "security_gates"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9cebff8542de10913e60e85b148459a72430de49"}
{"id": "train_45409", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "governance", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a0afb1b0f7d828aa7708e4ab2939bfcaa5aabc94"}
{"id": "train_45410", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "governance", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "78920b2b27bfe6fae6cffa2526204c02d186f0ac"}
{"id": "train_45411", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "repo_scale_reasoning", "tooling", "reproducibility"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3cffd4cf5f5c4895ca3ea6854f643480716a5930"}
{"id": "train_45412", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "repo_scale_reasoning", "tests_are_truth", "tooling"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b9c5dca6bc639927cb35b65aaa14ed292641cd10"}
{"id": "train_45413", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "ci_integration", "reproducibility", "security_gates"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d2b3c1fb17628ea819a697d66f9294a7d06680e1"}
{"id": "train_45414", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "documentation", "evaluation_metrics", "ci_integration"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c97c7296f4fb558d402799da5ec4e99acbd85ff1"}
{"id": "train_45415", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "ci_integration", "governance", "auditability"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f2f4ea26f5f2b24dbaf4f97a5839ea9f8ead2b76"}
{"id": "train_45416", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "security_gates", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "330abc29e723d1502b5ddbc9f1155f07a42e600e"}
{"id": "train_45417", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "ci_integration", "governance", "security_gates"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "07154649c872c9b6ad13a62b9107e23cd06cc955"}
{"id": "train_45418", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "ci_integration", "evaluation_metrics", "auditability"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7fbf76208875f8e3cc7a2c16f83ff46deb14a481"}
{"id": "train_45419", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "cost_latency_tradeoffs", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8f67e9790428c59c13fbe258e19a9ad7a33ae7ee"}
{"id": "train_45420", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "repo_scale_reasoning", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ddd34163e10626b15e34dad08bbf66b7fb4cce02"}
{"id": "train_45421", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "ci_integration", "tooling", "security_gates"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d92763bc72b91073f51052543f40176ef6deb0a2"}
{"id": "train_45422", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "cost_latency_tradeoffs", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3d9ebeac1cbaea7e9d85ca4267df6b7674bcd6c8"}
{"id": "train_45423", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "cost_latency_tradeoffs", "documentation", "security_gates"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d88454134419e7bdf33789dbc5a9e0c3bdce035c"}
{"id": "train_45424", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "governance", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0bf44260322d27e9c9403e0896d2521c3a5865fc"}
{"id": "train_45425", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "repo_scale_reasoning", "tooling", "reproducibility"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "767ac1923594a1f790bdc7cec8a45d5c39dab5aa"}
{"id": "train_45426", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "documentation", "security_gates", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "170306f94467c6f1223272f7b35942e7d08fe3c0"}
{"id": "train_45427", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "ci_integration", "security_gates", "evaluation_metrics"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ea32362cbb0969ae00084011f1c3552c08e1d2f5"}
{"id": "train_45428", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "governance", "tests_are_truth"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d84484d139c18b4d1bd82630f9cff9666ffcd9bb"}
{"id": "train_45429", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "governance", "documentation"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8d0f46a1aa999978ee95c131731359e227e4af75"}
{"id": "train_45430", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["governance", "security_gates", "reproducibility", "auditability"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "39addb71358d353d0237e7c5cf4e49fe97ee6e54"}
{"id": "train_45431", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "repo_scale_reasoning", "tests_are_truth", "tooling"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "889a51edcb34e0656885ea0451eec0f6343c0b42"}
{"id": "train_45432", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "tests_are_truth", "ci_integration", "auditability"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e4888a1ade2272b1a531f97cb9d1a325fa55cdda"}
{"id": "train_45433", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "cost_latency_tradeoffs", "security_gates", "auditability"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dfd2c6c574c9eab48736534308bbb389612a310f"}
{"id": "train_45434", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "auditability", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "059b678fed635ee528dba9c7e12c718c09554a47"}
{"id": "train_45435", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "documentation", "ci_integration", "tooling"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "adca71d176f6a4ff3fccf5a59d03e638c551c150"}
{"id": "train_45436", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0c83b29780bbc2e5f4c41f76898ba284506f9f59"}
{"id": "train_45437", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "reproducibility", "tests_are_truth"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e15f58c06ba9a6bbf593975bb659de1643f9d30f"}
{"id": "train_45438", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "tests_are_truth", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bb9c3bf4132ba94ac313e96c00acb93e5bb8c999"}
{"id": "train_45439", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "tooling", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "19a731ab8bd981f8756e3703f3ebcb1336158318"}
{"id": "train_45440", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "security_gates", "reproducibility", "governance"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "680bd029f388486557d1f4de9c02fffe8654f79f"}
{"id": "train_45441", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "evaluation_metrics", "tooling", "reproducibility"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fd1b1127d241e46b6274f42769b23797edcbed3d"}
{"id": "train_45442", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "ci_integration", "tooling", "auditability"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b14323814b70802425ae53c56b6c482e2f593aee"}
{"id": "train_45443", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "governance", "reproducibility", "security_gates"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a0300692b8eb8e4cf702460ae991d761b190dca5"}
{"id": "train_45444", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "auditability", "governance"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d3a3f3b2ff97fa6f8d590d6bf8eb8ee0a6e17841"}
{"id": "train_45445", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "cost_latency_tradeoffs", "tooling", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2dc71f86560434e88be57fbf25c2ccb7fba04b4c"}
{"id": "train_45446", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "auditability", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3312ab162ad54ea0323c2a071cd1afa20b0157e1"}
{"id": "train_45447", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "evaluation_metrics", "tests_are_truth", "security_gates"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7b825d1fa57259987c0a326d9df16747206bbb66"}
{"id": "train_45448", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "ci_integration", "governance", "reproducibility"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "00ad34a44f61ffbbe5ebe937791dd959d20b4209"}
{"id": "train_45449", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "security_gates", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "855fd810ffe59c02b166509d0e2abd6d5aad6062"}
{"id": "train_45450", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "tests_are_truth", "tooling", "reproducibility"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "53ccc119b3cde2e71cc00feaec48cb91d8d870c2"}
{"id": "train_45451", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "tests_are_truth", "tooling", "auditability"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4876660cf5bb16414b3d9d6c62032aae99de79c7"}
{"id": "train_45452", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "tooling", "security_gates", "documentation"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0ff16158788839e712049971203b062235ef4c89"}
{"id": "train_45453", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "security_gates", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bd4dd98229550397f7db58edf2d132f48f61d0fb"}
{"id": "train_45454", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "ci_integration", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b79531042cf95a55db5ad90bd49d8a642aed2686"}
{"id": "train_45455", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "auditability", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "865b909ed2e220041f356ecb5a046e156194d223"}
{"id": "train_45456", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "tests_are_truth", "reproducibility", "ci_integration"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e6ba995b99e0b73a48beb11d6c090ed0f614a5ca"}
{"id": "train_45457", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "reproducibility", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cdde8b5c36cb59ae1f692edadb3dd9a16b977780"}
{"id": "train_45458", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "governance", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "781c9994f2bdbe45739d6b6e0a7145a3d3ba0891"}
{"id": "train_45459", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "ci_integration", "documentation", "governance"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4c6079fdb7cb469b855c2c79125e03b684165c08"}
{"id": "train_45460", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "ci_integration", "security_gates", "reproducibility"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1087713866939887d947f06be605f1b69be33bbb"}
{"id": "train_45461", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "tooling", "security_gates", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "365ea6aeea9a99df971366a2b39779703d21e451"}
{"id": "train_45462", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "governance", "evaluation_metrics", "security_gates"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d494d2169b295f6dbf25c370c3e9ede534e9e5dc"}
{"id": "train_45463", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["governance", "auditability", "evaluation_metrics", "tooling"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c45071c31d39f7bec084a48276b42fa695636800"}
{"id": "train_45464", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "evaluation_metrics", "ci_integration", "security_gates"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b860f9bdea721c12de9f8d80720c2a78fb887075"}
{"id": "train_45465", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "tooling", "auditability"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ce0b8f38d460f94ca5246a6f24e41b0b1c5af70d"}
{"id": "train_45466", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "security_gates", "auditability", "reproducibility"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "35bf0bebae0d019e7c864519ea7666d5cdaa3d80"}
{"id": "train_45467", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "tooling", "documentation", "ci_integration"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bfe54c1f9d3e8160c8a35bcbcdee2223230f4bb3"}
{"id": "train_45468", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "evaluation_metrics", "tooling", "documentation"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ad22dbbd04fdbfdeb6ac521cb9c3a35075c81a7d"}
{"id": "train_45469", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "reproducibility", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cc6c5adf1df28ae0b5dc363fe7eea8187a2e46e2"}
{"id": "train_45470", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "reproducibility", "evaluation_metrics", "security_gates"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "84111df6673efa837be24e2b4060bf0ace3aa540"}
{"id": "train_45471", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "security_gates", "tests_are_truth", "ci_integration"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c018a5561d0c945fb2a6276effa39d074b71650b"}
{"id": "train_45472", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "evaluation_metrics", "repo_scale_reasoning", "tooling"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "15988a8a8144a2dd22da42a552ffe8248ef26b5e"}
{"id": "train_45473", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "tooling", "repo_scale_reasoning", "governance"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "19009fdb6f1b59f34c5eda5d4931a02e54b6499e"}
{"id": "train_45474", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "ci_integration", "security_gates", "tests_are_truth"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bcafe62777ab4de08e2b41836646228f4ddf592a"}
{"id": "train_45475", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "auditability", "tests_are_truth", "reproducibility"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c45397c1adcab884131d022ff1e346712a12c1e9"}
{"id": "train_45476", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "ci_integration", "governance", "auditability"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "84c23ad6a4d5614590f772a6cb9bcdca741c7a50"}
{"id": "train_45477", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "repo_scale_reasoning", "tooling"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5ad213088ebfa4db02c4b9df9d1c1397abd0a7b7"}
{"id": "train_45478", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "governance", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b146572cd26ea0f7cc87bee24a14b20f513ecbbb"}
{"id": "train_45479", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "documentation", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7d3ea98e925864bb10520d499f6d4f9c4f406bf5"}
{"id": "train_45480", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "evaluation_metrics", "tooling"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0ee20d4ab15c0f0f77b3f9d924e969745891661d"}
{"id": "train_45481", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "security_gates", "tooling", "documentation"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7b25224a2827f151d2e63b00f4a4bef618be25e7"}
{"id": "train_45482", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "documentation", "auditability", "evaluation_metrics"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f5f36341d9e03a421c3285570dd03f422bf6b0c7"}
{"id": "train_45483", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "reproducibility", "repo_scale_reasoning", "tooling"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cf580dd9d2084bd2be9c0b5d413267e3b954a097"}
{"id": "train_45484", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "tooling", "auditability", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cfea92171a9a66a699403489d2f7af16be102f45"}
{"id": "train_45485", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "security_gates", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9b9ace2188694902cef786f933c6df2af9a6ae6c"}
{"id": "train_45486", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "documentation", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "29ef928139b9b1ed0b86580657759d90777bde46"}
{"id": "train_45487", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "evaluation_metrics", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "79b52a69073c8335a9a59f36375a773fdf790221"}
{"id": "train_45488", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "reproducibility", "documentation", "auditability"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0118ade6aa42a4a569832f1426720bd1f8672ca5"}
{"id": "train_45489", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "evaluation_metrics", "ci_integration", "documentation"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d5f68335b03116b06f06237722cc93adcd401f4d"}
{"id": "train_45490", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "reproducibility", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1bb0cdb6c4226336fc96b9d21b6720758297cae1"}
{"id": "train_45491", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "governance", "security_gates", "tooling"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f41e29790975a860a248fa51c080fdec1729c0e0"}
{"id": "train_45492", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "evaluation_metrics", "tooling", "auditability"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "978ab70870cefbbc392980ab052f9c12af64d592"}
{"id": "train_45493", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "auditability", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b08391781014ebdb89f90be5ea39ae647d9af77d"}
{"id": "train_45494", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "repo_scale_reasoning", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c07ec15e4ac92435534c29e0533df1e6e37946eb"}
{"id": "train_45495", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "evaluation_metrics", "reproducibility", "ci_integration"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c98fe176a7ddb044fbc613675cee5064e2095746"}
{"id": "train_45496", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "governance", "security_gates", "tests_are_truth"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c6f90a393e4caa8b28cef784bf28b3a2c71e54c6"}
{"id": "train_45497", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "tests_are_truth", "repo_scale_reasoning", "documentation"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "af80e041f7f3e52d03a0963bb5f9cd106ddb3ae7"}
{"id": "train_45498", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "security_gates", "governance", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cc4a4ff6f04f47b1ab7371a958a5c8d37e184a64"}
{"id": "train_45499", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "auditability", "documentation", "evaluation_metrics"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a827f3c9a9c5bdb3775f0dba55b9876c48b37579"}
{"id": "train_45500", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "security_gates", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "10ca03207a71aa7486b1268df94b88242553748d"}
{"id": "train_45501", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "governance", "tooling", "tests_are_truth"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7b917e4c0e2ded63401d46eb7432c2c5ea58b49e"}
{"id": "train_45502", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "ci_integration", "documentation", "tooling"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0fa463b899706d660c4c73d875476c3f09f29176"}
{"id": "train_45503", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "auditability", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7d13a413377240baaf064bbb972eaca22268c175"}
{"id": "train_45504", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "tooling", "tests_are_truth", "ci_integration"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dd079c31a54ed325ca7ee3581dadb8c0518e83e5"}
{"id": "train_45505", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "documentation", "evaluation_metrics", "auditability"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9fd0ab7592067a370e546a13c8e0da058a00c648"}
{"id": "train_45506", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "reproducibility", "documentation"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7f4ea686677563d7f401b88ad483cf790761e3c0"}
{"id": "train_45507", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "evaluation_metrics", "cost_latency_tradeoffs", "governance"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "357289dedfa6831c79c6292d23e8bae32d210085"}
{"id": "train_45508", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "reproducibility", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2e6c8d148825a2112f57b6361f510ebcc290817d"}
{"id": "train_45509", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "documentation", "auditability", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "02d42b0b401d7350ed78b0571f42ca4dc53c3cfe"}
{"id": "train_45510", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "documentation", "tooling", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3c974aff7cd4e8cb756cc7f4f47fa2a35d1042d4"}
{"id": "train_45511", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "reproducibility", "governance", "documentation"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9e587a47e73b001f05ce29bded2cf1b210f4d1b6"}
{"id": "train_45512", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "auditability", "evaluation_metrics", "governance"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a0a0eb09212c1e3011aefe30a33495fadef8bfed"}
{"id": "train_45513", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "tooling", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fe4728006520a070c76d2ca936ab12e10407ad54"}
{"id": "train_45514", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "auditability", "documentation", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "96b3048b6ac0ef9f491310ec8b0ef70ce72a81b3"}
{"id": "train_45515", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "documentation", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f1e30eb72a5685667654a00ba50fe787c29a0f95"}
{"id": "train_45516", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "repo_scale_reasoning", "governance", "ci_integration"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8c6d0f24f888c9d234d5fb392579c0d0c9df2548"}
{"id": "train_45517", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "tooling", "auditability", "ci_integration"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "10db25b1e23163c93816e33359449476e7a728c7"}
{"id": "train_45518", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "auditability", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d77e84928022aec35ffc61ec4d6079e38c838f2f"}
{"id": "train_45519", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "tooling", "auditability", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5d6f57fe762d49922768b34ed2ca1c0f871b8f38"}
{"id": "train_45520", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "evaluation_metrics", "governance", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f2e8ba9c530c7c61832bfe184dd102d67307df96"}
{"id": "train_45521", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "tooling", "evaluation_metrics", "reproducibility"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5d69d2cc4bd464c8ca3bfa7417523bd9a1878e36"}
{"id": "train_45522", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "evaluation_metrics", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "adb00a00bdee7ffa40d2292a99312d3de1a08904"}
{"id": "train_45523", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "governance", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "78a7d8775c77cc08786c375924c115e207139377"}
{"id": "train_45524", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "cost_latency_tradeoffs", "tooling", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6ad6950d291469fa61453a1db678e46fc8d068c4"}
{"id": "train_45525", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "reproducibility", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4409a7ceeba3238b14e6a793378c6be70d931070"}
{"id": "train_45526", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["governance", "repo_scale_reasoning", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "78f4c22d6562fb5ecb5725bc20590f0af564abb8"}
{"id": "train_45527", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "documentation", "tests_are_truth", "reproducibility"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6afd780a4ed1bea7593b5da06b584d83f60c71b6"}
{"id": "train_45528", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "reproducibility", "ci_integration", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0b331d8ec0e67e26bda0ce22c8a77bc9702fc8bd"}
{"id": "train_45529", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "documentation", "reproducibility", "governance"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "59765da14ba33bd4d01c7cc09160a6188e303505"}
{"id": "train_45530", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "repo_scale_reasoning", "tests_are_truth", "reproducibility"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c316b06fdaf2ad507e2dd560aa2e1bbee642d899"}
{"id": "train_45531", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "documentation", "security_gates", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1d74b43229e173098fe4d0bbd13a1cf0555bc51d"}
{"id": "train_45532", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "documentation", "security_gates", "tests_are_truth"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4454a23b00e22b041419f4c728a926630a2809ef"}
{"id": "train_45533", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "evaluation_metrics", "security_gates", "reproducibility"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e726fad01eb296c3e670ab6f305f3bdf5f7b414a"}
{"id": "train_45534", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "documentation", "repo_scale_reasoning", "auditability"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ad4a46adc7e84a664dc2196d1d97d86504f7aff4"}
{"id": "train_45535", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "evaluation_metrics", "tests_are_truth", "governance"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a122310c146114cabd6c37c5ef81f6c84b5ff0b3"}
{"id": "train_45536", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "tooling", "security_gates"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8984656bb921cc2ab16352328f2869e7c2eeb71f"}
{"id": "train_45537", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "governance", "tests_are_truth", "documentation"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1a2206397cf5ec0489b033bbdc70bfe4feb28d11"}
{"id": "train_45538", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "repo_scale_reasoning", "auditability", "governance"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a7955c297e05708a04f93f1e346c3db426db6f01"}
{"id": "train_45539", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "auditability", "evaluation_metrics", "reproducibility"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9df6e1baa33ba82d0956c83e57e3cfb674efff88"}
{"id": "train_45540", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["governance", "cost_latency_tradeoffs", "documentation", "reproducibility"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0e31566574dd7a6e7c753959c2959274e9c70319"}
{"id": "train_45541", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "evaluation_metrics", "documentation", "tooling"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "249b30acd3754ccb4a0bf7a088ce6f5f5e160c68"}
{"id": "train_45542", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "reproducibility", "governance", "security_gates"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d46d43f4d697dcee82f2be1dee78524e15a7c46a"}
{"id": "train_45543", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "repo_scale_reasoning", "ci_integration", "auditability"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b1b0391d4849b7ea77be1e59a2930a9fd0980de8"}
{"id": "train_45544", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "governance", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ee3be44609c1be7e06fa1ba006ebff359d1bf56e"}
{"id": "train_45545", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "reproducibility", "tests_are_truth", "auditability"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "085642e87be9a435f1087afbfc26c1c0011b5df5"}
{"id": "train_45546", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "reproducibility", "evaluation_metrics", "tooling"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d39fd726e37eaa3fb985f89d4ff619c5b3d9f1e3"}
{"id": "train_45547", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "governance", "reproducibility", "tests_are_truth"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5069932daa1bc90f0d208b27af2e95e039cea86c"}
{"id": "train_45548", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "governance", "evaluation_metrics", "security_gates"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "90f10c94cbfce59346a4bc3cd677217f797ab7df"}
{"id": "train_45549", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "governance", "repo_scale_reasoning", "auditability"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "715d50865d361dd998086670353a1cf11f91f3ca"}
{"id": "train_45550", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "tests_are_truth", "tooling", "auditability"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e3c5e4fa45b2de53e101a0c95c53cde891554125"}
{"id": "train_45551", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "ci_integration", "auditability"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5097b112a46e957a605a5224d1d59fb17ddf0ccb"}
{"id": "train_45552", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "evaluation_metrics", "tooling", "reproducibility"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "40ff61dde4bff7fff8199d772a233596cf011e28"}
{"id": "train_45553", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "security_gates", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "61fb64aac84e82ee5e07942514fba90f76ea2b7e"}
{"id": "train_45554", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d830661d5fe51afce7963ad5bd084e2bffbc908d"}
{"id": "train_45555", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "governance", "auditability", "reproducibility"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0a571e55e066eb53ba2f6ea3bdb2ece7b2771388"}
{"id": "train_45556", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "reproducibility", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a48d2f334b8a97c81f0b12bf6f4992145347b198"}
{"id": "train_45557", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "tooling", "governance", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a3091126b0ab27965b8ade25706e32bcf715f59c"}
{"id": "train_45558", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "governance", "auditability", "reproducibility"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "058fd4989b5250009e70e95e8ca83d4c527bbb3e"}
{"id": "train_45559", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "governance", "reproducibility", "ci_integration"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0cd3c3480154022e42f29e1bf434e4ea0f74230a"}
{"id": "train_45560", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "reproducibility", "tests_are_truth", "tooling"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "804d4cb051d59705cd9a802898e22e8f13407ed9"}
{"id": "train_45561", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "governance", "tooling"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b5de638458dd199c66108c3468d04a8ae57004d5"}
{"id": "train_45562", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "ci_integration", "security_gates", "auditability"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4cb41868de41a4990a125354269d89cd3032edce"}
{"id": "train_45563", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "tests_are_truth", "documentation", "governance"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a876b74dd01011ecbeb163ebc158203e3f92c72a"}
{"id": "train_45564", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "security_gates", "evaluation_metrics", "reproducibility"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8bba1d55dc69575c4d8bc9e378ae442bca071ce5"}
{"id": "train_45565", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "ci_integration", "governance", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2e0505620b0a6a0c571477b911cfcb931ae7b99d"}
{"id": "train_45566", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "documentation", "tests_are_truth"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dd8ddf4ea5f29278ea5267aa5fe70de08e7da3c5"}
{"id": "train_45567", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["governance", "tooling", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c1d1e10ef17e57ae88f7297b3cbcbf8cac7d35bb"}
{"id": "train_45568", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "governance", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fd68868ed678ad6f3c6d043bfc28efd60631f2ff"}
{"id": "train_45569", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "tooling", "documentation", "auditability"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9e69b090b05d7ff9cfb19e4e0183a68c38e5efb1"}
{"id": "train_45570", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "auditability", "governance", "documentation"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7fbb395ec2fbd3c1af41632379d035c9914ab590"}
{"id": "train_45571", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "tooling", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "18b61a500304c8fb5ceeedd8d9bc86279f4e9ae3"}
{"id": "train_45572", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "tests_are_truth", "evaluation_metrics", "security_gates"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "40e05c60338aca11ced32fe1452d25fa71b1b69f"}
{"id": "train_45573", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "auditability", "ci_integration", "governance"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7b1ba48dbf2dea2a91a1a21f7f665a6dce7cd45f"}
{"id": "train_45574", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "governance", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a1277375e7cc57d931bad5582083f03cf184a873"}
{"id": "train_45575", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "governance", "security_gates", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4a567ff353286673ed99ec46c4659419a79d2ac1"}
{"id": "train_45576", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "repo_scale_reasoning", "documentation", "governance"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ad06a23af4b9925be759ecfd519796411c8c9694"}
{"id": "train_45577", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "documentation", "reproducibility", "tooling"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "98839e2634cada7c50404f3ef0206fdb128acc01"}
{"id": "train_45578", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "security_gates", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "533d4e0745ecd454008377e2b4886665df532714"}
{"id": "train_45579", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "reproducibility", "security_gates", "tooling"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "60099fab64bbeb6c1f966eb25b79d8dff5248479"}
{"id": "train_45580", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "governance", "auditability", "tooling"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c824859951bc8b25d20c3c8d904b0c9f16bd0876"}
{"id": "train_45581", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "auditability", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8f86329778b84dce9595c9aa8d5d5717450f54e6"}
{"id": "train_45582", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "evaluation_metrics", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "44f9ccfad59a967048dfe6354926c29dcdf797f6"}
{"id": "train_45583", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "ci_integration", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a12f13f1f11e92859e6af45aa5c14c5fed079260"}
{"id": "train_45584", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "evaluation_metrics", "tooling", "ci_integration"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "964251bcb2ec9dd3d931752262d0cb0264abd797"}
{"id": "train_45585", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "tooling", "security_gates", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8a16017f05c88a069f792e9610aa72282529189b"}
{"id": "train_45586", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "tests_are_truth", "tooling", "security_gates"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ff9011851087891713931bb00b9ab426b96422b1"}
{"id": "train_45587", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "evaluation_metrics", "auditability", "reproducibility"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f2c7332813ea7dd57755488262770d8d30ef616e"}
{"id": "train_45588", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "repo_scale_reasoning", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "546e1744ad7444806b8225d5d4ed4e925fb0f080"}
{"id": "train_45589", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "documentation", "tooling", "governance"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b7d35199d022253629a02a87a3ad5ba75e7548bd"}
{"id": "train_45590", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "repo_scale_reasoning", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "af17948f722df07f95b1a00281de27c758fd7a03"}
{"id": "train_45591", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "tooling", "auditability", "governance"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6b043890faeacbd0213e212ecc70abbf4d573fb3"}
{"id": "train_45592", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "ci_integration", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "96134be95f722789129fc500bab5a09ad25574a1"}
{"id": "train_45593", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "repo_scale_reasoning", "governance", "evaluation_metrics"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5cb6a59dd3fbdd6f7d46b0b578ad0a37dcdf66da"}
{"id": "train_45594", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "tooling", "evaluation_metrics"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c74ae1634f77f7f9ee697c8dff7594d20f688ae3"}
{"id": "train_45595", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "evaluation_metrics", "tests_are_truth", "documentation"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c9244cb540c0d776a70924bb9490aeca16699ccb"}
{"id": "train_45596", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "documentation", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "05a8f74fd7660813efe25145178302cfb7ddc30d"}
{"id": "train_45597", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "security_gates", "auditability", "reproducibility"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a0c3df38e8d3ffddfb9c4d5a0cadfb9ad5530a03"}
{"id": "train_45598", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "tooling", "security_gates", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6dbb4c3e6e23a7fb9b7d5f3217baf35f38a59449"}
{"id": "train_45599", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "reproducibility", "governance", "auditability"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7f9583abec7a9b05f63041b628fd4341ff5d940f"}
{"id": "train_45600", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "repo_scale_reasoning", "tooling"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "407b9fb72debba701439306af977e0718726e2b2"}
{"id": "train_45601", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "tests_are_truth", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3d2dc9c7f49e512cea4ac438300bd4b3fef34a0f"}
{"id": "train_45602", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "security_gates", "reproducibility", "tooling"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4a8edc969e135cbbe381c1516598c9add456b07c"}
{"id": "train_45603", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "reproducibility", "evaluation_metrics", "security_gates"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c6e8aacc123f4327e86a31cda802dd137045d268"}
{"id": "train_45604", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "tooling", "documentation", "reproducibility"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dd22fb91beedd626991fa006eb86b5f491efebe9"}
{"id": "train_45605", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "documentation", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d189466cb4029e394c33a666b184c828705c2c4b"}
{"id": "train_45606", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "security_gates", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d05e8aaf64abf7087b755bf0b0cae7c95dc255a5"}
{"id": "train_45607", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "auditability", "ci_integration", "documentation"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eaf6e755064f20d4dc09848a149eb28ad28107b5"}
{"id": "train_45608", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "repo_scale_reasoning", "tests_are_truth", "documentation"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fb8a37fd91d744a7bb38f7cf834855d58fed552d"}
{"id": "train_45609", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "tests_are_truth", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4ef4bcd09c82dfce99ee70f361422cccfbf3bbe8"}
{"id": "train_45610", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "reproducibility", "tooling", "governance"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "34b3b9570caf83ce03a48c110f4562a1fd498839"}
{"id": "train_45611", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "auditability", "ci_integration", "documentation"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b59d255fee841b72cebb5b43c7493c2bf68aa6d7"}
{"id": "train_45612", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "auditability", "documentation", "governance"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9687604bafe64fa885857b0e8253766b43d76b42"}
{"id": "train_45613", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "repo_scale_reasoning", "security_gates", "evaluation_metrics"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "83bac1bb11180a67e964acb2adad99034e217923"}
{"id": "train_45614", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "auditability", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0f42ddaa605a08d8845f49e5a4124cc32426d609"}
{"id": "train_45615", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "repo_scale_reasoning", "governance", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0a5aa840de96dbece231bf84ea462b7e67a6250f"}
{"id": "train_45616", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "security_gates", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d5a628d00cdc45c13e164868614aa35d3acaac6f"}
{"id": "train_45617", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "tooling", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7968ebfe77f2daa5250fe5c7816c4d1dbbeeb552"}
{"id": "train_45618", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "repo_scale_reasoning", "tooling", "auditability"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b5ba3c402012d729b328408f98e54847d9d86831"}
{"id": "train_45619", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "tests_are_truth", "auditability", "reproducibility"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "40e05c60338aca11ced32fe1452d25fa71b1b69f"}
{"id": "train_45620", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "repo_scale_reasoning", "reproducibility", "tooling"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "da305929da5fc82b6dc8a46be7fd140d267cfdea"}
{"id": "train_45621", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "auditability", "tooling"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "12e5d857cdc9c882fdaa260f15da3fa1e68b8acc"}
{"id": "train_45622", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "ci_integration", "governance", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3b64ad5b01977880f499992034e01bc7d87ed9d9"}
{"id": "train_45623", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "reproducibility", "security_gates", "tests_are_truth"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2f3044bff0682498280b7fe61609125b3c556d00"}
{"id": "train_45624", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "evaluation_metrics", "documentation", "security_gates"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "762ba32edcc8aab06608e478033f7e5d03f37a2e"}
{"id": "train_45625", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "repo_scale_reasoning", "ci_integration", "security_gates"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9166a02f06ff96b55a66885a4cfcbd0cba445ae7"}
{"id": "train_45626", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "governance", "auditability", "security_gates"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d289f915033d4e38e694376cd1161ef0ef16034d"}
{"id": "train_45627", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "auditability", "tooling", "evaluation_metrics"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b9673243d155f98246d1e91e3cbdd2580a4542b1"}
{"id": "train_45628", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "security_gates", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c471711779d0cb73dea3954973550fc0a9e7ca05"}
{"id": "train_45629", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "auditability", "tests_are_truth", "security_gates"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7e0eadbbb6b91bb69fe663ca569190f1b6e8cd1d"}
{"id": "train_45630", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "tests_are_truth", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "09a6392d56b8fc2a430951c25f7a484a7238e034"}
{"id": "train_45631", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "auditability", "documentation"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "34f7046f2af709a4a3a0f15aa0e031eddbe5c84b"}
{"id": "train_45632", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "governance", "reproducibility"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "368a8cdb845cc68398a389930dcf1f65c9957b6c"}
{"id": "train_45633", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "governance", "tooling", "evaluation_metrics"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ce6b4367462349d8e780a6118fd175a0c141f73b"}
{"id": "train_45634", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "security_gates", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "429020ec1bc9a558ea75c6f4fa49ae5dee2fff3a"}
{"id": "train_45635", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "documentation", "auditability", "reproducibility"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4876c8e749691ddcc4a95ec22b4aa9fb2398956c"}
{"id": "train_45636", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "reproducibility", "auditability"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c7334d2cbfc9af84bde4683bb78017fa099d0c80"}
{"id": "train_45637", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "tooling", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "53ff89cca5ebe3fd80cf669d899dc5ffae2870e8"}
{"id": "train_45638", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "tooling", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "60010612d9807e7c005b9849691a4a6d94341ada"}
{"id": "train_45639", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["governance", "reproducibility", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c0cf8dbba184c70066ddb781bbb700aee5d18610"}
{"id": "train_45640", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "reproducibility", "governance", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "908d0bae59d4c86e282536540ad0e6c2a0fa3e4c"}
{"id": "train_45641", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "tooling", "evaluation_metrics", "security_gates"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fd86ee6105ac3dbf2930b0864cc083ab30ac2432"}
{"id": "train_45642", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "auditability", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0f658bd4b28c5bd271c9fd6e372abb55ba2eefaf"}
{"id": "train_45643", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "ci_integration", "documentation", "governance"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "afb05e99f5e9b7badb7a14993db076eb8bd1d622"}
{"id": "train_45644", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "security_gates", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aea7608690b558345066a0a70cd3cfee3ddbf53d"}
{"id": "train_45645", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "auditability", "documentation", "evaluation_metrics"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2513012a0451f557045c31705e80dfe93da2f1ea"}
{"id": "train_45646", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "documentation", "reproducibility", "security_gates"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3f7d58ab5c7f67a3642d2753dc8b140d5a881522"}
{"id": "train_45647", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "tests_are_truth", "tooling"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f06cc432544446ce118cb99bb77cd6b73c67d9ca"}
{"id": "train_45648", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "ci_integration", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8f494462882bd7a915f590baf720ddc60bc9eae9"}
{"id": "train_45649", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "repo_scale_reasoning", "cost_latency_tradeoffs", "governance"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "34910bbc31a666a9e7023d780bbee03f1157b2ff"}
{"id": "train_45650", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "ci_integration", "tests_are_truth", "reproducibility"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8424ea94e1c318f4d8e3acd7d1e4bb5a465f3684"}
{"id": "train_45651", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "ci_integration", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ad981a4919ebe1c9e9cf20d96021f979c4a20464"}
{"id": "train_45652", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "security_gates", "documentation", "ci_integration"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ea7ccdcabe3079653fb6ce4714afeeb57dde2e1b"}
{"id": "train_45653", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "repo_scale_reasoning", "reproducibility", "auditability"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4194cc17c76059096ad81d098cd4fe4902145717"}
{"id": "train_45654", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "tests_are_truth", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1df8fecd705631b03dbefb5c1d400f23862b9bd2"}
{"id": "train_45655", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "security_gates", "evaluation_metrics", "auditability"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "93bb556f6ca8a1b74954d3e1a7fe68b8c90e3a28"}
{"id": "train_45656", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "evaluation_metrics", "governance"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d9f19113b3d0cc319b9a4226c9b20f35a8b6e700"}
{"id": "train_45657", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "governance", "documentation", "auditability"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1356476f063ba9efb9971401c566e348b2a7ccd7"}
{"id": "train_45658", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["governance", "ci_integration", "security_gates", "tooling"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f785220e091cf1c244d216868f172d257938f50f"}
{"id": "train_45659", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "auditability", "documentation", "tests_are_truth"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c5457934be61b8b8044e5281ba11e87fae1138b5"}
{"id": "train_45660", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "evaluation_metrics", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d6913d4d9b92a654f759de2bd1b7814c3b4b8b05"}
{"id": "train_45661", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "ci_integration", "reproducibility", "auditability"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dac5ecf28bed1fcdfcd5054584b195613f9758bb"}
{"id": "train_45662", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "tooling", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5ff740fd2acbfa4253e185c860fea3a7955637c8"}
{"id": "train_45663", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "reproducibility", "documentation"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6b69dfe7be2e69a4e54993346a66298006fb7228"}
{"id": "train_45664", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "tests_are_truth", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ed84b4a9879b194e3b6c15ed27c31c3eaa48099b"}
{"id": "train_45665", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "security_gates", "governance", "tests_are_truth"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cdaf4f130b6c41ac5b63b6c9cb7d011203074173"}
{"id": "train_45666", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "repo_scale_reasoning", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8d459b52ecbf47eb148dbd4196dece79066c81cf"}
{"id": "train_45667", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "ci_integration", "security_gates", "auditability"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "48e50893e1ec66574ffc6ee475b14cf1933561cc"}
{"id": "train_45668", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "evaluation_metrics", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1b28095bb517d0c0eda9f1c0bec3bcf400c1e239"}
{"id": "train_45669", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "auditability", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5eb28894b260372666bdc12703d51becaffccec7"}
{"id": "train_45670", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "documentation", "security_gates", "tests_are_truth"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "743d09369c084dd8f09624677a923aff0632c8b6"}
{"id": "train_45671", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "reproducibility", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2dd95b2255e79bc173d0ebb6d3120a38e674dcad"}
{"id": "train_45672", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "64cef56dd9004004f7009fbc46b05d9f07de902b"}
{"id": "train_45673", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "evaluation_metrics", "auditability", "tests_are_truth"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8dd86e9dfd639f5821f9340c21893a061c4de0d0"}
{"id": "train_45674", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "ci_integration", "evaluation_metrics", "governance"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4f9560f9953f1e15ce4ae13acc0ce785a4066a8a"}
{"id": "train_45675", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "repo_scale_reasoning", "auditability", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "00d01ba7d54f406f53131741d61bded26c6f4212"}
{"id": "train_45676", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "reproducibility", "tests_are_truth", "ci_integration"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d441b78ea706852e4eceb12c002225bf8354e808"}
{"id": "train_45677", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "documentation", "tooling", "ci_integration"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a93fbed7d1b780f53e38dbb4d3aa101e777ecac9"}
{"id": "train_45678", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "tests_are_truth", "governance", "tooling"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7010014d311694a54d7cce8930a7e79f3e6099ae"}
{"id": "train_45679", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "reproducibility", "tests_are_truth", "tooling"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b291485fd5ed8543ee8f95d083d8a7621fd685bb"}
{"id": "train_45680", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "security_gates", "evaluation_metrics", "tooling"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7250e95920e6b73d3b5736603c4535c054d5ee52"}
{"id": "train_45681", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "ci_integration", "reproducibility", "tooling"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2f8857281f2817af992b25c52b7e6cff668a32e4"}
{"id": "train_45682", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "ci_integration", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "01c5277e3cdc604c09cd1b85b5fcbe45fc649d54"}
{"id": "train_45683", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "documentation", "governance"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a66a2bb58f551fde26b0a63b4791bcbd98c989ce"}
{"id": "train_45684", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "security_gates", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "beccbe9b2dac710ff4e2d054c9135b687b148f8a"}
{"id": "train_45685", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "repo_scale_reasoning", "tooling", "tests_are_truth"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "80509e66eaa7c0af3a1ca7508d2fc24f22d0d43c"}
{"id": "train_45686", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "tooling", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4b48bace418723320cec5d8cac646de38f7a970f"}
{"id": "train_45687", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "reproducibility", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d125eaf328f26c33e2ccf1072e7237733261f596"}
{"id": "train_45688", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["governance", "cost_latency_tradeoffs", "ci_integration", "tooling"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2e2a7ea1afa4d8c98912d4ad1d1412683fb046c0"}
{"id": "train_45689", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "governance", "ci_integration", "reproducibility"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9fe48f90a9217d091b7315cb53b641b0adf4e382"}
{"id": "train_45690", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "governance", "auditability", "security_gates"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc59b00ed472f712a53c82eb3d63f6664747c72f"}
{"id": "train_45691", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "repo_scale_reasoning", "tests_are_truth", "auditability"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1d47d571773e2999817bdf1c7c0e67f1449091f0"}
{"id": "train_45692", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "governance", "evaluation_metrics", "auditability"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f53710748ed8858991c1bab159ad1119962d4f18"}
{"id": "train_45693", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "governance", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ad363dc07fb084cec9429b75ce0c702286a0c7d5"}
{"id": "train_45694", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "security_gates", "governance", "tooling"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0b51f4065cdb35c9aaa5ba208f4e2ac9378d601c"}
{"id": "train_45695", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "auditability", "documentation", "security_gates"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "daffb078574041307c1c0184c95f5de3d22abc07"}
{"id": "train_45696", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "tooling", "evaluation_metrics", "reproducibility"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7098555f96702cb4007c3275d102183f456cc94f"}
{"id": "train_45697", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "tests_are_truth", "evaluation_metrics", "reproducibility"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0578bcdbc7a030635cb558447e7a051f01f269b8"}
{"id": "train_45698", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "tests_are_truth", "governance", "security_gates"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aaeff21e7f34ac625c7b036eb2107d6b1399c69b"}
{"id": "train_45699", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "ci_integration", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b9a5fa06f23e65e106deeace7f7937750324d555"}
{"id": "train_45700", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "reproducibility", "ci_integration"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "da979a98780fca025defe7afa2c8bf9907bb3d66"}
{"id": "train_45701", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "reproducibility", "ci_integration"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8acd471390ed7b6737c35e325b548ebe7e4c98f2"}
{"id": "train_45702", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "auditability", "governance"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5503e99466b1698a760ff586e987822312b0bd5a"}
{"id": "train_45703", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "documentation", "auditability", "tooling"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "42294db117eb5ba876939dd3e8994b3275ee80d9"}
{"id": "train_45704", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "evaluation_metrics", "governance", "tooling"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "65297764e88bf46e30e735598fd4fdb2cd0c8044"}
{"id": "train_45705", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "cost_latency_tradeoffs", "ci_integration", "governance"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8bd4ea69306fbb4935e4ca06d96d7fbe90091432"}
{"id": "train_45706", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "documentation", "reproducibility", "auditability"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "da9c5fce130d8c35ef4ecc62c92189ed7cafe3f4"}
{"id": "train_45707", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "tooling", "repo_scale_reasoning", "security_gates"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2ebf110f3a4558dbdaef4517a0e8b0087182cc46"}
{"id": "train_45708", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "tooling", "ci_integration", "security_gates"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ec4080310099c11c21c2e3163b2246293362921e"}
{"id": "train_45709", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "ci_integration", "reproducibility"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fd73b69b29c77625c57636cd09485b007894a4e4"}
{"id": "train_45710", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "repo_scale_reasoning", "tests_are_truth", "documentation"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6c648264a18c4d132860602e2ac6d06ad1209f26"}
{"id": "train_45711", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "security_gates", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "790eb4374d2d7a234184d63d05af0a4a9d139600"}
{"id": "train_45712", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "reproducibility", "auditability", "evaluation_metrics"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7543a8ae8973a57de726d90ae36f763a5797de8c"}
{"id": "train_45713", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "ci_integration", "documentation", "evaluation_metrics"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ff2ede8739ca7cc0059cc65860d8f25d52791d9c"}
{"id": "train_45714", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "reproducibility", "documentation"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ea4fa7ee86f82e257c003fa0abd4bc6b4ab60ad3"}
{"id": "train_45715", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "ci_integration", "evaluation_metrics", "auditability"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0ad88948976a5bf9c92c48899e88768ca54dc14c"}
{"id": "train_45716", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "security_gates", "documentation", "tests_are_truth"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e04e2c29adc8813ca28cfc01220e7850d5bc83df"}
{"id": "train_45717", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "cost_latency_tradeoffs", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4d91024049c9c19f4275471de950d5c2bfcb11ab"}
{"id": "train_45718", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "tooling", "governance", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "139f5bff4a5875e47d8261d8b9ad22eb7b557e98"}
{"id": "train_45719", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "governance", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "15988a8a8144a2dd22da42a552ffe8248ef26b5e"}
{"id": "train_45720", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "tests_are_truth", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "71e29496cdba2867c2fe9ca0481ce2b9b06306b1"}
{"id": "train_45721", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "ci_integration", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5d0b0677e872bcb8e2371cf04b708240641e9dad"}
{"id": "train_45722", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "repo_scale_reasoning", "documentation", "tooling"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "832e81862bf341a4e15d76c13d7b78e74e3780ba"}
{"id": "train_45723", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "auditability", "reproducibility", "evaluation_metrics"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ee8bcd3fc93b3126407086afac6931611a66d631"}
{"id": "train_45724", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "documentation", "repo_scale_reasoning", "auditability"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8681fc1d7512a29887208e4c6c6930a65327afe0"}
{"id": "train_45725", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "reproducibility", "tooling", "ci_integration"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "61066ebeeaf48860a9b36159f1a8329be22e0d21"}
{"id": "train_45726", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "auditability", "cost_latency_tradeoffs", "governance"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ab75663ac324d21f43c5a5e885fb9d8ae99375ba"}
{"id": "train_45727", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "security_gates", "ci_integration"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e506a08a0c7ccfe55ab320de9c75d2fb2c25f333"}
{"id": "train_45728", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "tooling", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5f47a412b199c74fdc6aa1ef322bee4f07bf82e9"}
{"id": "train_45729", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "tooling", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aef985048ee27a820a8a2856c0e9db98b9ded104"}
{"id": "train_45730", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "tooling", "reproducibility"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "19050409bd386d9861514bc1ad70a98bd05e6e39"}
{"id": "train_45731", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "repo_scale_reasoning", "documentation"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "264aa07c9b9b6e5ee6471c3c5bf019a98ddcb0ee"}
{"id": "train_45732", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "evaluation_metrics", "tooling", "security_gates"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "64dd277cb4da199914f5aba7077fe1de5765a11a"}
{"id": "train_45733", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "tooling", "reproducibility", "auditability"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "85472f62bb83733a0b98b8bed7cee281bb43dbeb"}
{"id": "train_45734", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "evaluation_metrics", "auditability", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "42b15e3133690111e267af1ae7d9229c763912f2"}
{"id": "train_45735", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "security_gates", "documentation", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1cbdbb46984c0489627834d9b17d0adfefb30900"}
{"id": "train_45736", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "repo_scale_reasoning", "auditability"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "30e4a821a1e28ecac122d37d0b16f390b2b6499f"}
{"id": "train_45737", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "governance", "tests_are_truth"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4e88fe0416d75cb98847aaf20916e9abbdbf515f"}
{"id": "train_45738", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "auditability", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "df7997378c247e89f04b04136740f59f01062413"}
{"id": "train_45739", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "tests_are_truth", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fb0c6b9988fccb9365b3dfecc71c7bb857043e8e"}
{"id": "train_45740", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "reproducibility", "security_gates"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "78ed9aafa7c584a48e4f1ffeb97a3819b36d9e74"}
{"id": "train_45741", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "tooling", "security_gates"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fe7198dadb3b7c189b7362a30b6af1167aac4c62"}
{"id": "train_45742", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "evaluation_metrics", "tooling", "repo_scale_reasoning"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aa4593d51bd89e7502ee192335859267199f239c"}
{"id": "train_45743", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "ci_integration", "reproducibility", "security_gates"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "881bb9b5a3d2894eeb7c7f145a3ed519a3d7bd60"}
{"id": "train_45744", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "governance", "reproducibility"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "547dcf5c956afcea8d3dc2f097aab6b53357308f"}
{"id": "train_45745", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "documentation", "governance"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ac3116deb6e86abee1d50218eb0c3cf629d4ebf4"}
{"id": "train_45746", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "tooling", "auditability"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "07e254c5bbb167f12a4ac065ba1b2d3007d548d8"}
{"id": "train_45747", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "ci_integration", "security_gates", "tooling"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "193a2a7aeffff0b573cf72b4fc592209c36eb2dd"}
{"id": "train_45748", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "governance", "reproducibility", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e3bd48be3ec349ab7ccdd68a532bbd2aa75b9bcd"}
{"id": "train_45749", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "governance", "reproducibility"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0b5418be80632f6fb6aa5d619105c95b1e46220c"}
{"id": "train_45750", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "reproducibility", "ci_integration"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5dea96e7290c9b88ff63e1336c1ffc2111b6d4c8"}
{"id": "train_45751", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "auditability", "security_gates"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1192edd3be49d3873f7a74887e3b54c9d80e0b9e"}
{"id": "train_45752", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "tooling", "documentation", "ci_integration"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "782626e30157b4c396909e69ac833e1d255460b0"}
{"id": "train_45753", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "documentation", "ci_integration", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9fa56c4d40c8c7adca05be27c4fc1632a6671020"}
{"id": "train_45754", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "tooling", "documentation"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4617ac7f1948260168dbb7fcf38a5e971e163220"}
{"id": "train_45755", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "tooling", "cost_latency_tradeoffs", "governance"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "44ff8e52ccd6861d5db604acd2dded356603b0aa"}
{"id": "train_45756", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "ci_integration", "tests_are_truth", "governance"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eb6b8478e3b2dacc159dd2b73ba81e828254fdd8"}
{"id": "train_45757", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "security_gates", "documentation", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0c7ba8296b859eb1eaba854f4f5e2ca726ce66ca"}
{"id": "train_45758", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "ci_integration", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "790fc1a430034338f7fe24f195ff7210e514ed91"}
{"id": "train_45759", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "evaluation_metrics", "tests_are_truth", "documentation"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f5bf8c73f7c276cad46d3a22e8f67304e38c0852"}
{"id": "train_45760", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "evaluation_metrics", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9db6ddff0fe041263ad850dfbee2689d8bd59258"}
{"id": "train_45761", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "security_gates", "tests_are_truth", "auditability"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "310a6f83542ec275108bf8e6d490e2343efb3a87"}
{"id": "train_45762", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["governance", "repo_scale_reasoning", "evaluation_metrics", "documentation"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8c93ccfc98b1f01e970c3b7303347d5971d8b385"}
{"id": "train_45763", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "auditability", "reproducibility", "tooling"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cad0f127ec88f14602d71f89ea3100836826185b"}
{"id": "train_45764", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "tests_are_truth", "auditability", "documentation"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c063777bae659ea490f0bdde3ab177a5db85910b"}
{"id": "train_45765", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "repo_scale_reasoning", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8d9624ceaac84d4a9a40c2b791d4d7589f2a6b94"}
{"id": "train_45766", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "governance", "tests_are_truth", "security_gates"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b3a08f1dd50f3c377942a20314433fcfbabdb19f"}
{"id": "train_45767", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "documentation", "ci_integration", "governance"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e9ff8dec177463ab52f7e2355fe17b37ceb928ba"}
{"id": "train_45768", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "tooling", "repo_scale_reasoning", "documentation"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "59dda8f1fd0472f32d82d7e60644949f99c2cb06"}
{"id": "train_45769", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "documentation", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6fcd08507a6cf4583209be10af8271bb4e572e85"}
{"id": "train_45770", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "tests_are_truth", "security_gates", "governance"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0441179e0b10f8d30cfced57689c434b6941b078"}
{"id": "train_45771", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "governance", "evaluation_metrics", "documentation"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fad3e4e0fc3b78ed6256f537d04e533187792229"}
{"id": "train_45772", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "repo_scale_reasoning", "reproducibility", "documentation"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "73fafb56bff528f44c4820f0d4d05d7dc73dd89a"}
{"id": "train_45773", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "evaluation_metrics", "tooling", "reproducibility"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aa8043341e756dbb3845ae3f46779f9447b64b5c"}
{"id": "train_45774", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "governance", "tooling"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9fe8bd4c0b2fd3a82dc1f1a2d0bbc59b3e321222"}
{"id": "train_45775", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "governance", "documentation", "evaluation_metrics"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c0dfc7525da3d12fbefa36df45338e1c56105539"}
{"id": "train_45776", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "governance", "tooling"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "85088135c5208324318d8c80aed3337c1a11192d"}
{"id": "train_45777", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "security_gates", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a8cc6efb179dc3013bcff1be76a74921543ef5f5"}
{"id": "train_45778", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "ci_integration", "governance", "reproducibility"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "edc6cfcc4cc8ac7a6fb43bf90a61fa0cdf04e289"}
{"id": "train_45779", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "auditability", "tooling", "governance"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "adbcef64f2380b65e518fca5eedbf876d3c9e44a"}
{"id": "train_45780", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "auditability", "reproducibility", "documentation"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "039b71644af2d8cd8226f469a05c79d4d6d03fa8"}
{"id": "train_45781", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "ci_integration", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fdbd140b201b67f3df350acc3f246c751b6a9e68"}
{"id": "train_45782", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["governance", "reproducibility", "documentation", "tests_are_truth"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6311023c10bb7732935529ea66f292c479b8bf46"}
{"id": "train_45783", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "documentation", "auditability", "tooling"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "74301d46d7d4a54648e245649f715e90545695f5"}
{"id": "train_45784", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "repo_scale_reasoning", "documentation"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7aec61485d0f8d8ff638b5d56f88ae6666a75c02"}
{"id": "train_45785", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "evaluation_metrics", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8ad94c56f8f25543c8c097aab0a020d5720e70fb"}
{"id": "train_45786", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "evaluation_metrics", "reproducibility", "ci_integration"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "236fd6ea79917fe60a1f37a3e3d713c8cb14f7f3"}
{"id": "train_45787", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "reproducibility", "tooling", "governance"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a06d003b53451481a8b3ef43a35c93025662eb1d"}
{"id": "train_45788", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "auditability", "governance"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0978f0bef8473adab6a2636fbc719148bad6b960"}
{"id": "train_45789", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "evaluation_metrics", "auditability", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5f123ab618917bd1acd247493b07257afd64ee81"}
{"id": "train_45790", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "reproducibility", "evaluation_metrics", "auditability"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bcfdb8d7ab08340fec469705774f5564c9389778"}
{"id": "train_45791", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "reproducibility", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f6a13ed9c491b2427f3107bf2ef83cdde6093a3b"}
{"id": "train_45792", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "security_gates", "tooling", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e154c1dc7609f1eda62ea43220a8a56650c7008a"}
{"id": "train_45793", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "tests_are_truth", "auditability", "repo_scale_reasoning"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fee0b61306f43cb95b0b1eef7124f16de74a8191"}
{"id": "train_45794", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "auditability", "governance", "reproducibility"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f53d563a63df07894b7d5466ff3ca74bb167be94"}
{"id": "train_45795", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "cost_latency_tradeoffs", "reproducibility", "tooling"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "444a5436f207cc06a9831831f41e5bd194055f51"}
{"id": "train_45796", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "ci_integration", "tests_are_truth"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fa9fe2d9ce0ccef9e21634a2ec68658e817e8bcc"}
{"id": "train_45797", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "tooling", "auditability", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "704af66e9b627f50daf9cb81d9e30b9b2ff57ebd"}
{"id": "train_45798", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "evaluation_metrics", "governance", "reproducibility"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e691464c2582655d26f1db4e376cbf8722ad7b6b"}
{"id": "train_45799", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "reproducibility", "tooling"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1a79a529222be2069abe578589db2a69ee3e11bb"}
{"id": "train_45800", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "auditability", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9f61fe498703f4c7ddc2d9550a09beef690188c6"}
{"id": "train_45801", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "security_gates", "evaluation_metrics", "tooling"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "78bc6bafc3bb1ca9313365c176271fba830d3bd3"}
{"id": "train_45802", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "repo_scale_reasoning", "auditability"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b98c4e644f1de83a2c1c8104fae19003cb0a94c2"}
{"id": "train_45803", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "repo_scale_reasoning", "security_gates", "reproducibility"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3bf691969577b1e7266f58c849c3100d464cd566"}
{"id": "train_45804", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "auditability", "tests_are_truth", "reproducibility"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ac02e9d98b4d56d67d14a0b38ecc87957438e1e9"}
{"id": "train_45805", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "repo_scale_reasoning", "documentation", "auditability"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "49f62a78f766cff7a5ad7fdfcd69650e245a921b"}
{"id": "train_45806", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "security_gates", "tests_are_truth", "auditability"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e451a4f3b56baa760b52131d437b6a1bc05bbe36"}
{"id": "train_45807", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "tooling", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e6fb4ce3cf9acf3d8d5fd637371ff29061584454"}
{"id": "train_45808", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "repo_scale_reasoning", "auditability", "governance"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bcdedb35d564fefa815fd58c069208cebc3739ad"}
{"id": "train_45809", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "tests_are_truth", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "775c528dcec5838ce47db7a33ba8c70ec5efacf0"}
{"id": "train_45810", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "documentation", "tests_are_truth", "reproducibility"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "738d6062e4472cbd108d0ee1e4924b280d39472f"}
{"id": "train_45811", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "security_gates", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ee5a26d7815d2efa851f89c3692b4cbd2af94720"}
{"id": "train_45812", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "evaluation_metrics", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f6a13ed9c491b2427f3107bf2ef83cdde6093a3b"}
{"id": "train_45813", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "documentation", "ci_integration", "reproducibility"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f04b3160916a98ee9f09f7f08bda8645cfa022db"}
{"id": "train_45814", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "governance", "reproducibility", "tooling"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bd7c1adbcbaef5122182303fa63d74e9e28fa08b"}
{"id": "train_45815", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "tests_are_truth", "security_gates"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "df6bdae11dfdeb5cf260ac710af6c7ec6f82a73f"}
{"id": "train_45816", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "tooling", "reproducibility", "auditability"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "212cd3b3c0fe56ae2be6679e15bbe65f24084660"}
{"id": "train_45817", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "governance", "reproducibility", "ci_integration"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7e50de9444c0e304dedf40739be1ed74bae7e226"}
{"id": "train_45818", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "tests_are_truth", "governance", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c1e4e45311bd08537dc5efc8906c4a41b6552c96"}
{"id": "train_45819", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "evaluation_metrics", "security_gates", "reproducibility"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1e19c921636e1c77e45dc0d3169b2486f139d965"}
{"id": "train_45820", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "security_gates", "tooling", "tests_are_truth"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "de6758f6455d5fec2db0eea2981a25aa44a9f073"}
{"id": "train_45821", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "evaluation_metrics", "tooling"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8337b77ae9590b62d50dd3b6384d766a580678e8"}
{"id": "train_45822", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "cost_latency_tradeoffs", "repo_scale_reasoning", "tooling"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0ba0bc4afdf049f337cc149ab285ed4cc009d558"}
{"id": "train_45823", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "tests_are_truth", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "25cb99ff9b0aa115109194a1c13c9a478a4c3243"}
{"id": "train_45824", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "security_gates", "tooling", "governance"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b099243d1bf7df9e95b17776aed572736b66a7a4"}
{"id": "train_45825", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "tests_are_truth", "security_gates"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7857005340fb94de39f5ad336cf1cfe82d26ea37"}
{"id": "train_45826", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "repo_scale_reasoning", "tooling", "security_gates"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e19fb8a484417dc22a6f20d22bc2c02023363eb8"}
{"id": "train_45827", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "security_gates", "ci_integration", "tooling"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "402a63a5cef26541bcdada3d743711945920f100"}
{"id": "train_45828", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "governance", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cf580dd9d2084bd2be9c0b5d413267e3b954a097"}
{"id": "train_45829", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "reproducibility", "auditability", "security_gates"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ab65d3c12d179846334941ddfe8a65460a409626"}
{"id": "train_45830", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "security_gates", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c5ecebac2cab0654cf9eb6f492d4949f654ec599"}
{"id": "train_45831", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "cost_latency_tradeoffs", "auditability", "reproducibility"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "93e4751fbbd7062753a4ddbf8d003ee728ce5ee6"}
{"id": "train_45832", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "governance", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "09cb6624076032e5c275cfc100fb87ef9a313482"}
{"id": "train_45833", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "governance", "reproducibility"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1b090f4216ffb4db058b3da3938521610e3a31c7"}
{"id": "train_45834", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "ci_integration", "tooling", "governance"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1852f869dea93df0a0e64a59277212bdbf01dd97"}
{"id": "train_45835", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "ci_integration", "evaluation_metrics", "tooling"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "be54711122aff46037c01621a381086d16839ecb"}
{"id": "train_45836", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "repo_scale_reasoning", "governance", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ad29fa2549c09fcb18f32574166ee49c474e4673"}
{"id": "train_45837", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "cost_latency_tradeoffs", "reproducibility", "security_gates"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a1fbf16d9884e7a09188317eae5711987820256e"}
{"id": "train_45838", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "tests_are_truth", "security_gates", "ci_integration"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2e8bf23cbd50e2aee31433942f4d0f7846efb74b"}
{"id": "train_45839", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "security_gates", "reproducibility", "auditability"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8cea2f0ab13e461fc463ffaa8db4e025513ec92c"}
{"id": "train_45840", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "cost_latency_tradeoffs", "evaluation_metrics", "ci_integration"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "73516b8d55aa000e9fbb3564471a5cb992f508e9"}
{"id": "train_45841", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "governance", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6259155f769d9ac7330b5fffc72c1b254ec9bbe5"}
{"id": "train_45842", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "tooling", "documentation", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c803bc0b7da890046906ae6530784338d359311a"}
{"id": "train_45843", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "auditability", "governance", "tooling"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "016f2e27d541a4b56ed5171fe9fbbc211bd42e98"}
{"id": "train_45844", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "reproducibility", "tooling", "documentation"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "99a3dbd1dfc1bea93b0fa6c69f386762373f033f"}
{"id": "train_45845", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "tooling", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ce2458c384aa34ff6ed85c61fd0e0e2c08a740e1"}
{"id": "train_45846", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "repo_scale_reasoning", "ci_integration", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8a2a64fb0c5c4c56f4e04bed331413523a859536"}
{"id": "train_45847", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "evaluation_metrics", "tests_are_truth", "governance"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c0f91f48a23eab920ab1d3a31b49bf9a2c98ce37"}
{"id": "train_45848", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "ci_integration", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b5de638458dd199c66108c3468d04a8ae57004d5"}
{"id": "train_45849", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "governance", "tooling"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ff1d2656f0e5b6c0a8671eff9e3fa0fc752567fd"}
{"id": "train_45850", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "ci_integration", "repo_scale_reasoning", "tooling"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "497ed367a3dab335b8ccfbe5dedaf111d78902d9"}
{"id": "train_45851", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "governance", "evaluation_metrics"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6ccfc4ea83ef0f64927ec1e9947d6ffbae75c0f9"}
{"id": "train_45852", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "auditability", "ci_integration", "tests_are_truth"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "215a845a7fa5b99a81633047c6cc319c287b784f"}
{"id": "train_45853", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "tooling", "auditability", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "834fb0d0a713c1714a58907ba3eba1989d4c4c11"}
{"id": "train_45854", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "security_gates", "evaluation_metrics", "reproducibility"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "affa11772fe64c0352c9e2167a48caed609b5c6f"}
{"id": "train_45855", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "reproducibility", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c8f604e1ab9b261cab1271c8bb8bfcdf7d32213b"}
{"id": "train_45856", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "documentation", "reproducibility"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "89fdce024dcc0299dc05a03a0d25247dc93c183c"}
{"id": "train_45857", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "auditability", "security_gates", "ci_integration"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b376fe5f453450e80fd3eb7980cb573edf008332"}
{"id": "train_45858", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "reproducibility", "evaluation_metrics", "tooling"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ef4663e9b6d356cd7591fe00273c6cef35dc8eb3"}
{"id": "train_45859", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "security_gates", "documentation", "governance"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e451feaadd5b02a4cb1417664abf63536129f229"}
{"id": "train_45860", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "tooling", "governance"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "74d8fa29df6d7f7afe5bedfde7fb7d5903b15a7b"}
{"id": "train_45861", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "repo_scale_reasoning", "auditability", "security_gates"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "00b8c37535032a7546c3f78118a34a80a926e8e7"}
{"id": "train_45862", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "tooling", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "23acaf8db4a2174f68dd30346ccfda545e19bb4d"}
{"id": "train_45863", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "evaluation_metrics", "security_gates", "auditability"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b0fa75d582c151963516a06efa14afef53050075"}
{"id": "train_45864", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "reproducibility", "tests_are_truth", "security_gates"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "115eed8757156b2a2161ced46c8a18fe247f4b5f"}
{"id": "train_45865", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "security_gates", "ci_integration", "governance"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7c236ae2b18f19077279d2f58517f20bcf690876"}
{"id": "train_45866", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "governance", "reproducibility", "ci_integration"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d8a372f0cc48cf7544702d83eb60c66a809f85cc"}
{"id": "train_45867", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "auditability", "tooling", "security_gates"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "db9a864ad446fd6aaada1535a9264b33cc8ad28f"}
{"id": "train_45868", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "documentation", "security_gates"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "530e1480e1a55cf0be11672b081ca4491c811a7e"}
{"id": "train_45869", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "tooling", "ci_integration", "auditability"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "72bf759fb93ac00db06437933147be5f874353ff"}
{"id": "train_45870", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "documentation", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d494cdd7449479c6c4beb8877cb00ada7d91c2ea"}
{"id": "train_45871", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "documentation", "reproducibility", "security_gates"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5db201e96c2f83c6a8518e8888c18d9fc003a554"}
{"id": "train_45872", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "repo_scale_reasoning", "governance", "tooling"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4061492f1bda2f7f882afe0047e4c6bedca1ab19"}
{"id": "train_45873", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "tests_are_truth", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a50b4d8d1dcf4e3cfc90b0663ae665ebbc6823fa"}
{"id": "train_45874", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "ci_integration", "evaluation_metrics", "auditability"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "458cdec50480814995b2994c8b54df2836b5404a"}
{"id": "train_45875", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "auditability", "reproducibility", "tests_are_truth"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e8890185846755cb2c25dc1263791a8194df5067"}
{"id": "train_45876", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "tests_are_truth", "evaluation_metrics", "governance"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b013bbb18c29d13436738b7720fd53bdca30b2c0"}
{"id": "train_45877", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "evaluation_metrics", "auditability", "documentation"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7b1ba48dbf2dea2a91a1a21f7f665a6dce7cd45f"}
{"id": "train_45878", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "reproducibility", "security_gates", "documentation"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9a366af19be455fd8d31a2608d267a7a8d67f5e2"}
{"id": "train_45879", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "tests_are_truth", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3f847edad1af10f2e6e876647076ab70b88a5d6d"}
{"id": "train_45880", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "auditability", "documentation", "tests_are_truth"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4123cfa5c26525e847bddbaa866e0083a615af1b"}
{"id": "train_45881", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "auditability", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "19593c1195bf41473cdcf60bca41e20869a6784b"}
{"id": "train_45882", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "evaluation_metrics", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f790ce24fdd29fa1b32f35027c700d0bcd89064e"}
{"id": "train_45883", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "documentation", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ab73bc885f5e4da6372c385a6340344709e4d49c"}
{"id": "train_45884", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "auditability", "documentation", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4fe2e0d1985177a8034fbfde2af340c426650f95"}
{"id": "train_45885", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "governance", "security_gates", "documentation"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0eb8a06942b0211868595ddc62fff87bc210d941"}
{"id": "train_45886", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "governance", "reproducibility"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d494d2169b295f6dbf25c370c3e9ede534e9e5dc"}
{"id": "train_45887", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "tooling", "reproducibility", "documentation"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fe5d39819828f03342663ab457f22e446597aed2"}
{"id": "train_45888", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "documentation", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e0dac6ca89a26bb716bdef4bfaf2efdee80498ce"}
{"id": "train_45889", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "documentation", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c22ea591608735318d53ddd2dd4171f472093753"}
{"id": "train_45890", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "ci_integration", "tests_are_truth", "documentation"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5aefa14e1f775d6693fcb6a0299b1583a5c82730"}
{"id": "train_45891", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "security_gates", "evaluation_metrics", "ci_integration"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "606fa7718244f04ff0230e32b533ba336e0374b8"}
{"id": "train_45892", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "evaluation_metrics", "reproducibility", "tooling"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "99abcc1cac230846fb44439884e10af0c254d604"}
{"id": "train_45893", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "repo_scale_reasoning", "documentation", "tooling"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9b38df81d6e323c7b852da2a969ea859e5cdfb8c"}
{"id": "train_45894", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "evaluation_metrics", "documentation", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0630df47e7acbda6d0c3bc1ff76c14ec6ff69a11"}
{"id": "train_45895", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "evaluation_metrics", "ci_integration", "reproducibility"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eafc023107c7a786eaeb7ede6992f8be56828f67"}
{"id": "train_45896", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "security_gates", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ac9f014a33d759bae7403214caa385905eae1a71"}
{"id": "train_45897", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "auditability", "tests_are_truth"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "93ffdce60c745ea9124d6fd022ffa381b1fdf2ae"}
{"id": "train_45898", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "tooling", "documentation"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "35e46768fdbf6e4448f06f91428009e4430b0f42"}
{"id": "train_45899", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "auditability", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d008a2e469279df12a991eca8e30a8b6c28d40f1"}
{"id": "train_45900", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "governance", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "36d37027d62a1cf72ff2cd6c6c8273ef2fc91c83"}
{"id": "train_45901", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "auditability", "ci_integration", "security_gates"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0e84472e12e36586d5ba1045140fac4311aaf084"}
{"id": "train_45902", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "repo_scale_reasoning", "reproducibility", "tooling"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "53288cc75a99698bb12cbc0c336bd67c3a896886"}
{"id": "train_45903", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "ci_integration", "governance"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dafd628bb45541500f6d6caeca1fdd7cb9e24e0f"}
{"id": "train_45904", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "evaluation_metrics", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b74a44bbbe0fe1c29a018d2eb343b04cecc7e94f"}
{"id": "train_45905", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "auditability", "security_gates", "documentation"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e241e4e54420704eff078523b7862d081abf6e7b"}
{"id": "train_45906", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "governance", "documentation", "tooling"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b565c64d0c4a70d116134a166ed3ca30a9a709ee"}
{"id": "train_45907", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "auditability", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "873c174b40a5cfe5bfc0c930c8dde9a1c6dbb405"}
{"id": "train_45908", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "documentation", "reproducibility", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "13a6cb0a7a8ba5e0d6cae9d2675a99d621978212"}
{"id": "train_45909", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "documentation", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "344f4edcf074e00dd568377c7882a4ad6c4e71f6"}
{"id": "train_45910", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["governance", "repo_scale_reasoning", "auditability", "evaluation_metrics"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3eda36850b8fa5020b95cc1cd9f12a23536b8639"}
{"id": "train_45911", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "documentation", "ci_integration", "auditability"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3c4d673780ac9349498c5848f88056ba700a73c1"}
{"id": "train_45912", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "ci_integration", "reproducibility", "governance"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "57c38ec0d62d59d82d625c84f9355b7de20f46d8"}
{"id": "train_45913", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "evaluation_metrics", "documentation"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d57ef54b77b9abc2bda82bc02be12dd3c4e106a8"}
{"id": "train_45914", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "reproducibility", "auditability", "ci_integration"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "15f5707cc454ddb7a01bd43effcf1c15bf9ded64"}
{"id": "train_45915", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "security_gates", "tests_are_truth", "reproducibility"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dc7c872c0b6df6b3d3857027c5080fb325df8bbf"}
{"id": "train_45916", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "evaluation_metrics", "cost_latency_tradeoffs", "governance"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "47c0aa877c7a838242dc3371a593837cc41c72fd"}
{"id": "train_45917", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "ci_integration", "security_gates"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "873c59c57962e446afac7f64bc2d515888c11d01"}
{"id": "train_45918", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "auditability", "security_gates", "evaluation_metrics"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bed6bdb0c11a5ef93b23e7ac6c2cfb37222e0008"}
{"id": "train_45919", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "ci_integration", "evaluation_metrics", "governance"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d11301a105adc727e13935aa8ff11a856c0b63f5"}
{"id": "train_45920", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "tooling", "documentation"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8212290f9ff89abd4342ce07cd00d7e1938440ec"}
{"id": "train_45921", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "reproducibility", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bfe54c1f9d3e8160c8a35bcbcdee2223230f4bb3"}
{"id": "train_45922", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "ci_integration", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f46288e9a505f6766774f18988031329e49863e3"}
{"id": "train_45923", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "security_gates", "tooling", "documentation"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e60a641ca130837a8025326e7acafa36bfe2bb4b"}
{"id": "train_45924", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "governance", "tooling", "security_gates"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0b5b85f5054deaee8de11d7dd098cb0069066d88"}
{"id": "train_45925", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "repo_scale_reasoning", "tooling", "auditability"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2bc3dc832569a02d782f50a462648c09cf80db2c"}
{"id": "train_45926", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["governance", "cost_latency_tradeoffs", "security_gates", "auditability"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "93ffdce60c745ea9124d6fd022ffa381b1fdf2ae"}
{"id": "train_45927", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "evaluation_metrics", "governance", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "333b6cf1b269b3c4a2e8a0a2ed3b0cd1a2117f20"}
{"id": "train_45928", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "security_gates", "auditability", "documentation"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "888f7c1cbc2a27a53a4aa9d09e43edd992b3a4fa"}
{"id": "train_45929", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "cost_latency_tradeoffs", "security_gates", "tests_are_truth"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7411deb0a500b85515aae27ac5bb68b40fdeadcd"}
{"id": "train_45930", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["governance", "documentation", "tests_are_truth", "ci_integration"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a166653188fb6bbb6ed663f16790f220fac3dda3"}
{"id": "train_45931", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "auditability", "tooling", "repo_scale_reasoning"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6ae02f636092caad7599ee87ded0b49603c81dc2"}
{"id": "train_45932", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "ci_integration", "reproducibility", "auditability"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b7a9342baedf7cf7b9ad666c0d717c626b6ebca9"}
{"id": "train_45933", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "reproducibility", "ci_integration", "security_gates"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ef791fc2022727e3b1c9996c5cc996b32e59c18c"}
{"id": "train_45934", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "documentation", "tests_are_truth", "ci_integration"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cba8ca861365a66fe64d5877b2dde6c2312642fb"}
{"id": "train_45935", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7eca5bd768663ef2f408d44d60de8d0e18eac0e2"}
{"id": "train_45936", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "reproducibility", "governance", "evaluation_metrics"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "27677e025a29ab9aebd616f4107d57e4ccd5202b"}
{"id": "train_45937", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "documentation", "reproducibility"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e26fc156a2ece3feb6306b5eb23de11653bd4f96"}
{"id": "train_45938", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "evaluation_metrics", "documentation", "tooling"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "42a17cbe15f840612a0387d2bcdbba4e9fbcfb9d"}
{"id": "train_45939", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "documentation", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a2bd3748165dfa64312ee77f2819a3c58ffe7143"}
{"id": "train_45940", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "documentation", "auditability"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dd0a8be38449b3cd4c655c7979f9735ae0aabdeb"}
{"id": "train_45941", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "documentation", "repo_scale_reasoning", "governance"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b1ac6c0b89c36400037259c7168245f7358c4a79"}
{"id": "train_45942", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "auditability", "cost_latency_tradeoffs", "governance"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ad7fdc3c28c30cf1e77e8be319c4a959fd412cfa"}
{"id": "train_45943", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "security_gates", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "53775c701fd87520d1dcbd54a46e721ba9971672"}
{"id": "train_45944", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "reproducibility", "tooling", "governance"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3eda36850b8fa5020b95cc1cd9f12a23536b8639"}
{"id": "train_45945", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "repo_scale_reasoning", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "70b9e869f7de21d638251ca3e37da908b2bfcef1"}
{"id": "train_45946", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "reproducibility", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "840d7eb46c90aa7c7e973b98cf6e824771861fed"}
{"id": "train_45947", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "tests_are_truth", "auditability", "tooling"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "de2c9f64e218fb1f40cd7a6b5b9e9b7eba1e2163"}
{"id": "train_45948", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "evaluation_metrics", "governance", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "97a9d00133ebe0dd803f0a66e8f61b724abac2b8"}
{"id": "train_45949", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "tests_are_truth", "auditability", "tooling"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7b3dc66b1839e26d350b61cf36a79fa2b3a0d893"}
{"id": "train_45950", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "reproducibility", "tests_are_truth", "auditability"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "72e786c3ee906a7a4d7a5ecdbc0b08e60ef44cee"}
{"id": "train_45951", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "evaluation_metrics", "tests_are_truth", "auditability"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a0f3a71aa0aea6242ac151de11d2d18b5ee0fc69"}
{"id": "train_45952", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "tests_are_truth", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0c84b1ffde1cf8bcce7d7497b8f92ded8d50a468"}
{"id": "train_45953", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "evaluation_metrics", "auditability", "tests_are_truth"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8747185a7f5d61f21905c1e239542135ba6ed9b2"}
{"id": "train_45954", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "tooling", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "42b5cd24e946832da2cdcc6e3170e5c4851602a6"}
{"id": "train_45955", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "security_gates", "repo_scale_reasoning", "documentation"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "825e4c254fcc635367b6d2fcceb85337ed90c5b8"}
{"id": "train_45956", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "tests_are_truth", "reproducibility", "tooling"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1e7a638578e998cecc5d04b87b84a125df9e17be"}
{"id": "train_45957", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "documentation", "auditability", "tooling"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "98905552fa64ed883f10dae5376ac86485af1fc0"}
{"id": "train_45958", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "repo_scale_reasoning", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ae6e829973e052a5776f6a0c11c6ed7aef5d70df"}
{"id": "train_45959", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "documentation", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1c828ef0f00b426f7e08fcfcae2b479332294e62"}
{"id": "train_45960", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "evaluation_metrics", "ci_integration", "reproducibility"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4967f9c5c562d570a2d6362ee4ebeb10326d8851"}
{"id": "train_45961", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "security_gates", "evaluation_metrics"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8d2e203c10288c2992cbc29de070acfaef77e0d1"}
{"id": "train_45962", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "tooling", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bdfeb5c29f2416438e1f5fcfbb779a7c9ec84fb2"}
{"id": "train_45963", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "governance", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8bc5c9159e76b5fbe7f7911902f5311d1b8791b9"}
{"id": "train_45964", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "auditability", "governance", "ci_integration"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "abb0b01bcb15bb9b4e126a15036fc70968fc858c"}
{"id": "train_45965", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "governance", "reproducibility", "tooling"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "85b2a3da3f6f0e0efd3ba7b5ae245c0b0d6537ed"}
{"id": "train_45966", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "reproducibility", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dec4185e24f66bdbaab61e6ec3790a6809c8698f"}
{"id": "train_45967", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "reproducibility", "ci_integration"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "41aeac108cf06fd99db80cad36e737824d26af98"}
{"id": "train_45968", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "ci_integration", "repo_scale_reasoning", "auditability"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "770e5eff2e67ff12e07b946787b1bc0b92e0128c"}
{"id": "train_45969", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "governance", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0ba2ef7d41ab9cf2b0f8ac2246f036b713d4d036"}
{"id": "train_45970", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "auditability", "security_gates", "tooling"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f47eca1707b9700951d060c8b9aa79e6f784150a"}
{"id": "train_45971", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "repo_scale_reasoning", "tooling", "security_gates"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "630e0cf864bfb93dbc3491dedb0319742a6fba8a"}
{"id": "train_45972", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "evaluation_metrics", "repo_scale_reasoning", "documentation"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "33bdc23be6076b9f9a2a10571dc6130144e3e544"}
{"id": "train_45973", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "tooling", "reproducibility", "governance"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f86945452faf2a9a3088d831c4a9f2f6fb2994fa"}
{"id": "train_45974", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "security_gates", "reproducibility", "tests_are_truth"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "655d9fcd55ac7a254feb7c7f32367d05b6151dda"}
{"id": "train_45975", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "security_gates", "documentation", "governance"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f628efd90c74ce95eab660fb38db3500aa3b1769"}
{"id": "train_45976", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "auditability", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "61be3fec5a75db53b7cc4a384cd400f0dd8e97fa"}
{"id": "train_45977", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "auditability", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "38b370b3e801c4c712e753d4759ea4bdd44ca112"}
{"id": "train_45978", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "governance", "ci_integration", "auditability"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6aecb1d7174b088f8e96d7719934b6287b2f62be"}
{"id": "train_45979", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "auditability", "evaluation_metrics", "ci_integration"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "224fe3414a2140523a7347870285719259ef4324"}
{"id": "train_45980", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "governance", "reproducibility", "evaluation_metrics"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5f33f620d1a8e721f813bfc0abfa1d21add0ebba"}
{"id": "train_45981", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "tests_are_truth", "auditability", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c98ac729f5403de13ee616a8a20bc06a4edb1657"}
{"id": "train_45982", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "evaluation_metrics", "documentation", "tests_are_truth"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5a4d7f25f0cc46af75086dad1e7bc9e6255cf4d3"}
{"id": "train_45983", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "documentation", "tooling", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f78c0f6990faeb2501d059e14b7c6ed25e666ed3"}
{"id": "train_45984", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "auditability", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7cd81a5b60b59ddc08cc99e2cd9778e8a9ea8442"}
{"id": "train_45985", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "reproducibility", "documentation", "evaluation_metrics"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6e7006f6ed3e1261b4544bee9f41692132a57a32"}
{"id": "train_45986", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "repo_scale_reasoning", "auditability", "security_gates"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dbd11852ea51e78e0d0e3a75449aca47c64e4c05"}
{"id": "train_45987", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "documentation", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e5e845325cdb4f1650c50e213790630a2b8ca6c5"}
{"id": "train_45988", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "auditability", "governance", "security_gates"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a51d2c4ad264325b307dfa797e3f5ade7d8616fd"}
{"id": "train_45989", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "governance", "auditability", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "28f59dc82777d0f28395c679d40d158cc7be80dc"}
{"id": "train_45990", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "governance", "reproducibility", "ci_integration"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b8c177696d70c81b23e9b0c24f51fc115f023d97"}
{"id": "train_45991", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "tests_are_truth", "repo_scale_reasoning", "documentation"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b37c50c499cda72d8844bc3ec5e8ad319ae58872"}
{"id": "train_45992", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "governance", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b3f614984ec3bf8b56e1b93c53a5bb433eecbd9c"}
{"id": "train_45993", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "tests_are_truth", "cost_latency_tradeoffs", "governance"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "907baf22d69990a88f6bb82094d6f017cb71e8f6"}
{"id": "train_45994", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "auditability", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b493fb340ccfc2605f8c7d0cc803318dc05c212f"}
{"id": "train_45995", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "tests_are_truth", "documentation"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "58f922ea7e8ab74835175e5687836082910ad410"}
{"id": "train_45996", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "documentation", "tooling", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "102a6642efa770721e97bc6bb65ccc7c789630cd"}
{"id": "train_45997", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "reproducibility", "tests_are_truth", "auditability"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "14223fa02bf3803285a5a000d2fd931c3f96207c"}
{"id": "train_45998", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "auditability", "tooling"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c8c1e72a2a4a2aed55324ef6e385a210b6e7e564"}
{"id": "train_45999", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "reproducibility", "tooling", "security_gates"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "20e14ab0532a12f0ea4e37022a239f613e185dcb"}
{"id": "train_46000", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "auditability", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "137ac4c8989fa60dbc00c3b9858b13dc49aef689"}
{"id": "train_46001", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "tooling", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5abfd26217ad208f707415e2f673ca5c0fb24494"}
{"id": "train_46002", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "tests_are_truth", "repo_scale_reasoning", "auditability"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "364c5f8ff755b0fb792637d536a27f2ccf99fbda"}
{"id": "train_46003", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "governance", "security_gates", "reproducibility"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1d8c147951f59ddd206404bd8749f7db297c2fa8"}
{"id": "train_46004", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "tooling", "auditability"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "79e7cc056c5e9761fc0ba50ab41a08655d9d31a3"}
{"id": "train_46005", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "reproducibility", "documentation"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2f3a6ea36355a03f3684a29e14a53819630d9eeb"}
{"id": "train_46006", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "reproducibility", "governance", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ba81a406c83f0e21a6b5c0a9110cfe4aa9e3f5c0"}
{"id": "train_46007", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "auditability", "security_gates", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a672464ebc3691689f36f61f265db67a2b04348c"}
{"id": "train_46008", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "ci_integration", "documentation", "security_gates"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2fbd1490babd4386083f7566254084450c291b94"}
{"id": "train_46009", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "auditability", "tooling", "governance"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2bdb6d7c83ce1770b0608cbd5d3bc3e2b3f55f6d"}
{"id": "train_46010", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "tooling", "documentation", "auditability"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c476d147ead669d64679c8cd2aa66f554291abb6"}
{"id": "train_46011", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "evaluation_metrics", "governance", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1192edd3be49d3873f7a74887e3b54c9d80e0b9e"}
{"id": "train_46012", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "governance", "documentation", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4d6f669eb9ae148417aa7f452c7d0c3d0878ed21"}
{"id": "train_46013", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "reproducibility", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2556f98fbe06ef53a41e878a9a7599d69ed16f63"}
{"id": "train_46014", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "auditability", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8ca6df4f0cf342c0ec54af62aa1e61e3dba53f55"}
{"id": "train_46015", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "tests_are_truth", "evaluation_metrics", "tooling"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "abacfe585000362bd0103aedcda8418689483c93"}
{"id": "train_46016", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "tests_are_truth", "governance", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "af7ce911cd079db4d68cefe4a9f5207366dd6b4f"}
{"id": "train_46017", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "ci_integration", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "735e5b25be1b4953c525779778bb9314664a5bb1"}
{"id": "train_46018", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "repo_scale_reasoning", "auditability", "documentation"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "25fc3166ef5e14907d2e7b1caa8bc500c8c0e2f9"}
{"id": "train_46019", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "security_gates", "tooling", "tests_are_truth"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6f1b0822b3ab47d89c5a64be3f28479c1ecd923c"}
{"id": "train_46020", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "reproducibility", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "76a46615dcc12079f79e617b651d1a619c0611a4"}
{"id": "train_46021", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "evaluation_metrics", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b26bb55dcdb4b0c2cf85a5eb4df41cf6fff6947a"}
{"id": "train_46022", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "tests_are_truth", "auditability", "evaluation_metrics"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "34941ab4efdb8b5fe48d42dfc74e0402457f6ae3"}
{"id": "train_46023", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "repo_scale_reasoning", "auditability", "documentation"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "97f95447c275faeb1c4b6c422d144b489904160c"}
{"id": "train_46024", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "evaluation_metrics", "documentation", "security_gates"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "128fdde0c755bb5155d38a57f11e8175d8b62a45"}
{"id": "train_46025", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "governance", "tooling", "evaluation_metrics"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "365ea6aeea9a99df971366a2b39779703d21e451"}
{"id": "train_46026", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "ci_integration", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c41d2f44a0d5f8eec8ce0fe1864831bfd4be248c"}
{"id": "train_46027", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "tests_are_truth", "ci_integration", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "adcd674dd2f8e8146c64a814931df57129109e67"}
{"id": "train_46028", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "cost_latency_tradeoffs", "auditability", "ci_integration"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8291588bda31045f8ff9ae83761ca48621e6ff0b"}
{"id": "train_46029", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "85ff536d3f5b2bbcfce3796b57596b5c50684016"}
{"id": "train_46030", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "security_gates", "documentation"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eaf57b6f71597ae486643557838fa24fa60c29d5"}
{"id": "train_46031", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "repo_scale_reasoning", "auditability", "documentation"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f309d769211b31ce082fbe9b7bc12e3c7f34da7b"}
{"id": "train_46032", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "auditability", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0c5483d89b61021f96bc0a67cb8a9f16babef0d4"}
{"id": "train_46033", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "security_gates", "tooling", "documentation"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "015a5ae8e239b7fd9b69c6bf5cef8df703795ec7"}
{"id": "train_46034", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "tooling", "auditability", "governance"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0c4ccdfebc7360c6033f20a48ce0080d01c968bd"}
{"id": "train_46035", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "repo_scale_reasoning", "ci_integration", "governance"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ae36195831f0319cb381ef03e65156f9652dc619"}
{"id": "train_46036", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "cost_latency_tradeoffs", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc628b513931eb68b8d4a51b66c6ddf592b54574"}
{"id": "train_46037", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "documentation", "auditability", "evaluation_metrics"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7e016805e7a8acc3d8d7f2dfbc8ccaecaaf77120"}
{"id": "train_46038", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "evaluation_metrics", "reproducibility"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1d160eb2a024642969d549fc0a59e6ae02a943d5"}
{"id": "train_46039", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "tests_are_truth", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0b51f4065cdb35c9aaa5ba208f4e2ac9378d601c"}
{"id": "train_46040", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "ci_integration", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fcddeae68574cc57640c4e3c3f772cda7103a735"}
{"id": "train_46041", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "repo_scale_reasoning", "governance", "auditability"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9c927cafaa15686c965761eeee2dc03879d6a213"}
{"id": "train_46042", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "auditability", "reproducibility", "ci_integration"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6ad15cc3eeb3cda2c220530a8634429eaa0a2d13"}
{"id": "train_46043", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "security_gates", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ecafb6cd0f25506751884792e634b5fe977aabd6"}
{"id": "train_46044", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "documentation", "tests_are_truth", "ci_integration"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "03ba8e448ae8db4c92e6e38bc9a9b21ba83c8244"}
{"id": "train_46045", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "tooling", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6d7af6cd57def1ca399e9da51c981ca02e630896"}
{"id": "train_46046", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "security_gates", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9692da5ea434dfbe4ad7ebfcaf62589345dc086d"}
{"id": "train_46047", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "security_gates", "auditability"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "36baa7a3a058292d82a1c2175384b4af53b9b6a0"}
{"id": "train_46048", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "ci_integration", "governance", "tests_are_truth"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b264c661a517c1635a375d608ef0b61e8e56ef4f"}
{"id": "train_46049", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "auditability", "repo_scale_reasoning", "security_gates"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f1f4dc72e824b5732c24f74ce10c3ee09d2e3e69"}
{"id": "train_46050", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "repo_scale_reasoning", "governance", "auditability"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f2a6a5de0ca59d863a8529d3d619fcc02453c156"}
{"id": "train_46051", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "reproducibility", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2456b66f7bfc4ab5ede829d8675134149fb5fa1e"}
{"id": "train_46052", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "ci_integration", "reproducibility", "auditability"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "535b1f435d1b2109366c548bcdb295feca6dda95"}
{"id": "train_46053", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "documentation", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b31ca137de3f070013a786fd059bdd8a9fea450c"}
{"id": "train_46054", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "cost_latency_tradeoffs", "governance", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8304a18d128b25f724c3b006113574afa08fb066"}
{"id": "train_46055", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "reproducibility", "evaluation_metrics", "ci_integration"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1affda88b1bb83af2b0fbc83a774b9e3d199bbf2"}
{"id": "train_46056", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "reproducibility", "repo_scale_reasoning", "documentation"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "887a8eff45f926ba2620c5d1f350ea779a47d588"}
{"id": "train_46057", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "tests_are_truth", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0b331d8ec0e67e26bda0ce22c8a77bc9702fc8bd"}
{"id": "train_46058", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "ci_integration", "reproducibility", "auditability"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "170aee58e1c2b7ea00350ca1b1adc1d27af25a0f"}
{"id": "train_46059", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "repo_scale_reasoning", "evaluation_metrics", "governance"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b260e93be82774d7e1c75d70573c275a12032d29"}
{"id": "train_46060", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "repo_scale_reasoning", "security_gates", "governance"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "19b21cbc017e06971677ca915c4a50e8e5aef0cc"}
{"id": "train_46061", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "reproducibility", "evaluation_metrics", "auditability"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "94fa0cdbc025fd331ebe1afa23ee81c191789fc8"}
{"id": "train_46062", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "documentation", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "23287131987a6d30a5e8119318d81045a4e749d6"}
{"id": "train_46063", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "repo_scale_reasoning", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cc41061f1ad8a28d2443b59a9655679d93a7d6cf"}
{"id": "train_46064", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "ci_integration", "tests_are_truth", "reproducibility"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3af8654b7dbfcaa12c4841c7e732b82e5c789e55"}
{"id": "train_46065", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "auditability", "tests_are_truth", "ci_integration"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d0be4e3c9be0f5c0afed56c677792887af132732"}
{"id": "train_46066", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "documentation", "repo_scale_reasoning", "tooling"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "febb7ef05b611dcd8a7d3a9ea535b4295254b402"}
{"id": "train_46067", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "documentation", "repo_scale_reasoning", "tooling"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6eba96f7df8bd1a7724d20672504f1f0d9f95ecb"}
{"id": "train_46068", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "governance", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2b74eeafe7b6dfe0526f071fc04ea1e1e82be15e"}
{"id": "train_46069", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "auditability", "cost_latency_tradeoffs", "governance"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "55e7003b115a0acf58dec68f75d1d67a929894e9"}
{"id": "train_46070", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "auditability", "evaluation_metrics", "documentation"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d3ef700992a90cee47d336786f953b7ae26ffcef"}
{"id": "train_46071", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "reproducibility", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "50353f345efb3b158cd8965ce5aa4345ffd1b799"}
{"id": "train_46072", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "security_gates", "tests_are_truth", "governance"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b97edd173f1589b226cbaf1e231c364947739ad3"}
{"id": "train_46073", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "documentation", "ci_integration", "security_gates"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9ce17238cb4e00d87940a904e4d5646f0b416a25"}
{"id": "train_46074", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "tooling", "governance", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a005aa6d157043a91b5e646f13d5712e54829d7a"}
{"id": "train_46075", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["governance", "repo_scale_reasoning", "ci_integration", "documentation"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dde6f5b469778d120b340c488cf34301a03e1b39"}
{"id": "train_46076", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "auditability", "tooling", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "648230042ff698ded340c930140a9d7025774b6f"}
{"id": "train_46077", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "security_gates", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ddc01bdff018057adc844470db89bbf4307d3864"}
{"id": "train_46078", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["governance", "tooling", "documentation", "evaluation_metrics"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "36e298bbd54eeb526fe01f1ca26c5a42f5714a7b"}
{"id": "train_46079", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "governance", "tests_are_truth", "tooling"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "25fedcb43c39d99c038d0f3bb17f6204f6ccb2fa"}
{"id": "train_46080", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "evaluation_metrics", "tooling"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "267920ee4bfc6786cb7588bbd588b2947046afea"}
{"id": "train_46081", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "documentation", "governance"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c02ce0a13237ee0133e4ed462d05e4ffd4735d25"}
{"id": "train_46082", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "ci_integration", "cost_latency_tradeoffs", "governance"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "37cdeab78224b7c9054c7dbcd742d2dd06a3090e"}
{"id": "train_46083", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "security_gates", "auditability", "reproducibility"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "efdbe375940aad0b7f7d86a7e45b5d8a80a22270"}
{"id": "train_46084", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "repo_scale_reasoning", "tooling"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "004d973028054599dcb198025d3cc32bff043f0a"}
{"id": "train_46085", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "auditability", "documentation"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "92dd7a24e3a679d722c4fbac469c33a49b69eab5"}
{"id": "train_46086", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "security_gates", "documentation", "ci_integration"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "895b1e1cc947347695d8e694ba91181d79630187"}
{"id": "train_46087", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "tests_are_truth", "governance", "tooling"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "48d33959b0e7d343329e3834f63fe1fbbd9211e4"}
{"id": "train_46088", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "cost_latency_tradeoffs", "security_gates", "reproducibility"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "82165521e798a6c74ba1c6bba49acd35b5b95b2f"}
{"id": "train_46089", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "documentation", "security_gates"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "465b93556b8086c19cb3134c507ebda74aee5b9c"}
{"id": "train_46090", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "cost_latency_tradeoffs", "ci_integration", "evaluation_metrics"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c63d575bd4a1964abd89535bb821c3207a8da73f"}
{"id": "train_46091", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "reproducibility", "security_gates", "tooling"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "237de752299c301edf147815f0c27abfe713d474"}
{"id": "train_46092", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "governance", "auditability"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7cb8962e619f9ad28f49225debc6cbc7d295b492"}
{"id": "train_46093", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "documentation", "auditability", "tests_are_truth"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "60c4fd2820193af11e6c4e26ff6c0137cd234f5c"}
{"id": "train_46094", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "documentation", "governance", "tests_are_truth"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7c7c1cb7e2ec5ab0fb8d023e4dd64215fec46125"}
{"id": "train_46095", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "auditability", "security_gates"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9c9c0be4af001d199df74f825aa290f022bdf280"}
{"id": "train_46096", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "reproducibility", "auditability", "security_gates"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e28c0d2cb2b5cb07b0be46c2b6d3d0f1ace66a17"}
{"id": "train_46097", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "repo_scale_reasoning", "security_gates", "evaluation_metrics"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "468e58d610f2c02734f11cc01e3605f300b8b0f0"}
{"id": "train_46098", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "tooling", "evaluation_metrics", "ci_integration"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "304db5a7c62fe821b5d309dd872b4247ee70fc25"}
{"id": "train_46099", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "tests_are_truth", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3063fc61a2125f38bd7f10e50d74c18d947b368b"}
{"id": "train_46100", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "documentation", "security_gates"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "087254dde96c1041b1cee4d92b082a18471c7078"}
{"id": "train_46101", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "reproducibility", "ci_integration", "governance"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4aef6afcd8395a42d22e72f46af1aac04e88bc0f"}
{"id": "train_46102", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "auditability", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3d8399e6fafdfa7f6504baf108cca2610c03c109"}
{"id": "train_46103", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "documentation", "auditability", "reproducibility"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f20020deeff1692a2bb2d69c225a90e22b583743"}
{"id": "train_46104", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "security_gates", "reproducibility", "tests_are_truth"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0c4ccdfebc7360c6033f20a48ce0080d01c968bd"}
{"id": "train_46105", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "security_gates", "auditability", "tests_are_truth"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6311023c10bb7732935529ea66f292c479b8bf46"}
{"id": "train_46106", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "reproducibility", "auditability", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f25c699f167be89155a248dff77b0d4e9b89babd"}
{"id": "train_46107", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "documentation", "governance", "ci_integration"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "65d4ce49368bc0a34050524cadd631f80ce6b4d5"}
{"id": "train_46108", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "evaluation_metrics", "ci_integration"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e68622844137a07f46e846e241b9f3f22c531c52"}
{"id": "train_46109", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "documentation", "security_gates", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b090e561f9b6ff8a7ab9c2a35e335d7e211d2350"}
{"id": "train_46110", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "auditability", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "34f1400640ed7bff47fa0bed98f468a6d9d1964c"}
{"id": "train_46111", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "tests_are_truth", "security_gates", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b0b0b4267186cb27bca531b19bd837d970b302dc"}
{"id": "train_46112", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "auditability", "governance", "documentation"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "31f13250e4d97545747dbdee1ab9a99a0fce4c14"}
{"id": "train_46113", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "repo_scale_reasoning", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0118ade6aa42a4a569832f1426720bd1f8672ca5"}
{"id": "train_46114", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "documentation", "reproducibility", "ci_integration"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4dab624fa552ebbf1115d5781055806be174263b"}
{"id": "train_46115", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "documentation", "evaluation_metrics"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3ef523b4b30c563165e0b9cec9f1564f71c9ef74"}
{"id": "train_46116", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "ci_integration", "documentation", "auditability"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2233141afe5b15ba1d7228c5207055d7cef58ab6"}
{"id": "train_46117", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "ci_integration", "auditability", "reproducibility"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a29983b15055410a9092f9267f34abe0ff2ab40b"}
{"id": "train_46118", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["governance", "tooling", "tests_are_truth", "documentation"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0c6fe787b255e0e329548c0b5fc0d97339f63e63"}
{"id": "train_46119", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "reproducibility", "governance"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "afc9946ba38105fbf9d571ab24ef8831bd28ac05"}
{"id": "train_46120", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "tests_are_truth", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "554a165ebd8fd2e44f095a041e8e77beea5df4ae"}
{"id": "train_46121", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "governance", "repo_scale_reasoning", "security_gates"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "105855ea1a857bb2f9985be87c02899539bfd614"}
{"id": "train_46122", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "auditability", "ci_integration", "security_gates"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "94c5f3c11e966398cef2928d5ec2ca23adc51924"}
{"id": "train_46123", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "repo_scale_reasoning", "tooling", "tests_are_truth"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8995dcc1ee03419e336fde101587245c7ee402ab"}
{"id": "train_46124", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "governance", "security_gates"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fd87c33edf00523798ee8bd7d34a57ea0629a5dc"}
{"id": "train_46125", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "documentation", "governance", "evaluation_metrics"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f9eae8f8e534d22d5ce259815d8b60007f4eef72"}
{"id": "train_46126", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "governance", "auditability", "tooling"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ca95c104b673603ea8ba4752f2f037e59489b521"}
{"id": "train_46127", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "ci_integration", "repo_scale_reasoning", "auditability"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cdcb4ee342c5833f1d54c9d1e9c2bb12afcfa88e"}
{"id": "train_46128", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "repo_scale_reasoning", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c0ebe1699ba47bc53a6f730a2663fb609ceee97b"}
{"id": "train_46129", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "reproducibility", "governance", "security_gates"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d416d3a1b65c3a946e3298dbc2ad0043da1c09c5"}
{"id": "train_46130", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "ci_integration", "auditability", "evaluation_metrics"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fc2485ae7eaf55b2bcef59c724b90ba059085711"}
{"id": "train_46131", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "ci_integration", "governance", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0dc53eb02fee8bbe2238a15c94eb91869642900e"}
{"id": "train_46132", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "tooling", "evaluation_metrics", "ci_integration"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "492c12e8a64e2de8ea4c56ced652485430c01f9f"}
{"id": "train_46133", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "security_gates", "tooling", "reproducibility"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "db1d2d6b4f1534185ba9bbbc303b0ba633648c50"}
{"id": "train_46134", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "tests_are_truth", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d3c55836b5e7d8ad6c524a7350b534d6caf67df4"}
{"id": "train_46135", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "tests_are_truth", "documentation", "security_gates"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "689f6e288216779816983d7ed9a291783ba35d7b"}
{"id": "train_46136", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "auditability", "security_gates", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "053ffc2cd05af22c7391d8bf3073910ec3fa54ec"}
{"id": "train_46137", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "tests_are_truth", "auditability", "ci_integration"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9f924935010f376e2ca28ec2cb6fdfee788e2b18"}
{"id": "train_46138", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "reproducibility", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9c87b53dc83ac7b9ddf4638aab7712973d8236cc"}
{"id": "train_46139", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "ci_integration", "evaluation_metrics", "security_gates"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "970a4cbfcf5dd91b2d5e6348b704c5ea0752d0d7"}
{"id": "train_46140", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "tests_are_truth", "documentation", "governance"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "097fdd55ccf4a444eabe77b9a86cd73bf76c6c92"}
{"id": "train_46141", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "ci_integration", "governance", "tests_are_truth"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b1b97b698d8e7274fa49a4a0d5a966c1590965bb"}
{"id": "train_46142", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "ci_integration", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "81bb7a103dae3389c40eddcb00780b510161bb47"}
{"id": "train_46143", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "governance", "evaluation_metrics", "tooling"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9f2498590d9f58e36deded2a4919190bc2e60c88"}
{"id": "train_46144", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["governance", "security_gates", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "23454cef1da059af0261c868c4516eba6635687c"}
{"id": "train_46145", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "security_gates", "cost_latency_tradeoffs", "governance"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "972e45e41c3a1b3886f608b967696026a26c76f8"}
{"id": "train_46146", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "reproducibility", "tooling", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c3ed57862f5cbb68d06551bd1d6283e4a523a2bb"}
{"id": "train_46147", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "ci_integration", "governance", "auditability"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e6ba995b99e0b73a48beb11d6c090ed0f614a5ca"}
{"id": "train_46148", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "auditability", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ca603e6d323a6f30a0f0ade6a535a812f4315d21"}
{"id": "train_46149", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "ci_integration", "evaluation_metrics", "governance"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ed4dbf94e228f0243f0f90bf01a46f035c5c5dd5"}
{"id": "train_46150", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "tests_are_truth", "ci_integration", "auditability"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c42799de40890a2c58b293fce160de0ede98da31"}
{"id": "train_46151", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "tooling", "evaluation_metrics", "reproducibility"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3f34430a52cf3d075e27e4640f33a6d94feb8987"}
{"id": "train_46152", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "documentation", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dec4185e24f66bdbaab61e6ec3790a6809c8698f"}
{"id": "train_46153", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "security_gates", "documentation", "auditability"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ee72848198820c0a7617ad9e9ebe402bdc49af7b"}
{"id": "train_46154", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "documentation", "auditability", "tests_are_truth"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b05037d5a7a20ac0963583682be30bd150e3fa94"}
{"id": "train_46155", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "reproducibility", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5f381c013799ee331a8e5ace2ec9fe013ab5f9e3"}
{"id": "train_46156", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "governance", "ci_integration"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ec1434b9bc5b83c5c42d9a70ec11ed01516700f7"}
{"id": "train_46157", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "auditability", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "87361510ee0d276453712ae0171d9c640cbb29e9"}
{"id": "train_46158", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "tooling", "reproducibility", "ci_integration"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "19beb05cbc89a2b3b0b8a2503ba0e06e0f4b92f8"}
{"id": "train_46159", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "ci_integration", "tooling"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7594fa8fcd18fbef21c4e6da5e9e9822a98c79e6"}
{"id": "train_46160", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "evaluation_metrics", "auditability"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f6d37913cdbc908f4f4bfe4235f7e60bdf061528"}
{"id": "train_46161", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "governance", "security_gates", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8a8b103e030e4b7cbb1b007a207c0cebd07418cf"}
{"id": "train_46162", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "reproducibility", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6ef878f2db1c4a0fffa1f2cfd3672c4a9d3f16e2"}
{"id": "train_46163", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "repo_scale_reasoning", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a792d03410c4cead080bbb1e04ef18783335caf4"}
{"id": "train_46164", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "reproducibility", "security_gates", "tooling"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d15d919bc01adfeb7fd6140156e2d92d2e0c1b7e"}
{"id": "train_46165", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "tooling", "repo_scale_reasoning", "documentation"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b3708d9ea50ddc19ac9889af380b53ca52e07d7d"}
{"id": "train_46166", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "documentation", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c90cc3a29a8897a7e3e4e305a3657ea1fd4577c8"}
{"id": "train_46167", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "reproducibility", "repo_scale_reasoning", "auditability"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6acb897f06cd7989d00bf8ea2b6654c195957434"}
{"id": "train_46168", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "evaluation_metrics", "auditability", "tests_are_truth"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ec5bf6e59dda649e20fbe00910d19c9b1850affd"}
{"id": "train_46169", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "evaluation_metrics", "auditability", "tests_are_truth"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5ba0f6c469fff8a9ba1903c163f83216e6df93c5"}
{"id": "train_46170", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "security_gates", "tests_are_truth"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "379a5a36060b134454624db7f03f7b1d13f3bd77"}
{"id": "train_46171", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "cost_latency_tradeoffs", "tests_are_truth", "governance"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a805c85028bec6667371ac8864f597831a9043fe"}
{"id": "train_46172", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "reproducibility", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4befb2c4376feaaad5418298be9fde7d3eba1ca0"}
{"id": "train_46173", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "ci_integration", "evaluation_metrics", "auditability"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "07f974458f6c216d9fc94910b4ebc2a6ba9ed9e6"}
{"id": "train_46174", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "evaluation_metrics", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0ab3f3a2bb172bc06ca4659823854060505fd799"}
{"id": "train_46175", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "ci_integration", "governance", "security_gates"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e6c145cef6418ee8f71c83374951789b20d8f380"}
{"id": "train_46176", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "auditability", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f9c74de10c606c16c5af90f45d889cc01b8e28a6"}
{"id": "train_46177", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "repo_scale_reasoning", "tests_are_truth", "reproducibility"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b0bfbb4708d2c3794d76878e78a1a8631ad71725"}
{"id": "train_46178", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "evaluation_metrics", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "06389f9aa3fba1e0f44c954f8c6c231d7b42d1e5"}
{"id": "train_46179", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "tests_are_truth", "tooling"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3a628556eded50f88c9d445f5ab70d7e239143e3"}
{"id": "train_46180", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "cost_latency_tradeoffs", "security_gates", "tests_are_truth"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aac2ac08171e5392c7f006dca3cffc87e92512df"}
{"id": "train_46181", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "ci_integration", "reproducibility", "auditability"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "24eba2181ca53ae193d6babf60c7a42f9c106c1d"}
{"id": "train_46182", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "tooling", "auditability", "governance"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fd378791a7e9777151f27ccf260b29c330e4558c"}
{"id": "train_46183", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["governance", "auditability", "tests_are_truth", "ci_integration"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "71286830105408c1d919981949a32f8cc0b38b66"}
{"id": "train_46184", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "ci_integration", "governance", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1a842f29ad8b1dee40b94dedfc1cbe085ceb784e"}
{"id": "train_46185", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "governance", "auditability", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0bf44260322d27e9c9403e0896d2521c3a5865fc"}
{"id": "train_46186", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "reproducibility", "repo_scale_reasoning", "auditability"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "339ecdbc2ca38fdc9c300b8ce8dbb8fe7feac0c9"}
{"id": "train_46187", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "auditability", "evaluation_metrics"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5db201e96c2f83c6a8518e8888c18d9fc003a554"}
{"id": "train_46188", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "reproducibility", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e19fb8a484417dc22a6f20d22bc2c02023363eb8"}
{"id": "train_46189", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "reproducibility", "governance", "ci_integration"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a5681f585f0c7f643ede68d01e2fd4397659e6c4"}
{"id": "train_46190", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "governance", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "73516b8d55aa000e9fbb3564471a5cb992f508e9"}
{"id": "train_46191", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "tests_are_truth", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "57956b6cb191933e6e19c1faa71c3d2285d13275"}
{"id": "train_46192", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "tooling", "reproducibility"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8cae9d31ced657075915115580fd63f87334b6f6"}
{"id": "train_46193", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e3e545e17779e715ac43b71dce543dbcbf8d17bf"}
{"id": "train_46194", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "security_gates", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5cf4e04e5f18fd953e69d380ddf8a75c140fa0b7"}
{"id": "train_46195", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "documentation", "tooling", "ci_integration"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eabb92298cf294c0163fb3c52e12ca700ab5afaa"}
{"id": "train_46196", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "tests_are_truth", "documentation", "governance"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "66af19d0525f39a31153686a7444afe06e43c089"}
{"id": "train_46197", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "ci_integration", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc40f0819c01acb810bfc7624748b3d7d409e365"}
{"id": "train_46198", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "ci_integration", "auditability", "tooling"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f5daff8397bbba8b1341919348eaa0446f801277"}
{"id": "train_46199", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "reproducibility", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0b3df58113dd703fa27807f38685f3f7aaf13a59"}
{"id": "train_46200", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "evaluation_metrics", "governance"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9d008e18e21e62705d8398da94f47dad02bd5705"}
{"id": "train_46201", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "reproducibility", "documentation", "tooling"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aaa959ad798b512bbcc47e8f22c4db27ff206b79"}
{"id": "train_46202", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "reproducibility", "governance", "auditability"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "37677db34224740be552276e556b95c6f39a4972"}
{"id": "train_46203", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "tests_are_truth", "ci_integration"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "535e1430a648b7e7e42bbc7c1eb057f66e0b8d54"}
{"id": "train_46204", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "auditability", "ci_integration", "documentation"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6b4b8dc4e94bcd9b09f60040e20c1ca6b7129644"}
{"id": "train_46205", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "governance", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e47e3c8b282b6bbfcd3d8368d3b1326629421fc0"}
{"id": "train_46206", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "ci_integration", "governance", "auditability"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c3be35998aa1177c74ed0ca3ba3834c2c18a45e4"}
{"id": "train_46207", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "documentation", "auditability", "tooling"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c702830f27a0f4b564406364989b9fb9e201ae83"}
{"id": "train_46208", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "evaluation_metrics", "tooling"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2c64f1dfddd4f4aafe06154e45af5c855dc6fcb3"}
{"id": "train_46209", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2878dd91d5796bc9b1a17ef72e6b5e963ee21754"}
{"id": "train_46210", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "reproducibility", "tooling", "tests_are_truth"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bb61456e5266a7d285539fe9d2196c690c142a18"}
{"id": "train_46211", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "documentation", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e87e5c2c62e6da63c25efaee820308eeca338342"}
{"id": "train_46212", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "cost_latency_tradeoffs", "evaluation_metrics", "reproducibility"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc15de94fe3b7c29ab193d75a2d10b7e79e3974c"}
{"id": "train_46213", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "tests_are_truth", "documentation", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6259155f769d9ac7330b5fffc72c1b254ec9bbe5"}
{"id": "train_46214", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "governance", "reproducibility", "security_gates"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "398b79fca1d43396f8b7169869c483e0c8449f8f"}
{"id": "train_46215", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "tests_are_truth", "ci_integration", "governance"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dfdc02ca850dad0c269e6c96386fef74c3c02404"}
{"id": "train_46216", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "ci_integration", "tooling", "governance"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cb50ef30768ec9a3c3032d2d1ea5e4f92b336e5f"}
{"id": "train_46217", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "evaluation_metrics", "security_gates", "documentation"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "76aee81f85de901dee74a7c6ceb7e5bee00f910c"}
{"id": "train_46218", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "auditability", "tests_are_truth", "tooling"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3410bdf4813ccd73c3a15fd3fa499faec3640ef4"}
{"id": "train_46219", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "tooling", "documentation", "auditability"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fd7c8c753efd7360856465252e9dac6eaa950753"}
{"id": "train_46220", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "evaluation_metrics", "tooling", "tests_are_truth"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3e41738627729729a37441c013bde4ac182d0778"}
{"id": "train_46221", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "tooling", "evaluation_metrics", "security_gates"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8b485a1ae5a641ff37f6ce4cc386f07468ff6cac"}
{"id": "train_46222", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "reproducibility", "documentation"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8bdd6a277ce236a8659e585f0364815b194e4d71"}
{"id": "train_46223", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "repo_scale_reasoning", "reproducibility", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "63667a6a749b967300c96f5e2b905536fa8a977b"}
{"id": "train_46224", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "ci_integration", "governance", "documentation"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f12e288c8a4e4eb9a57690527f7f1fe993fba301"}
{"id": "train_46225", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "98076009faa296a81c64ab62e7b7412151c71c64"}
{"id": "train_46226", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "ci_integration", "tests_are_truth", "auditability"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "385f86c0a51dbac2b9ce9245a62e640e3cc528d5"}
{"id": "train_46227", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "auditability", "documentation", "ci_integration"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "01935252c47e813c99458b96a97698bc2b95debf"}
{"id": "train_46228", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "documentation", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "87c6a2b977b8c8173193b735032dd35e3f240c7b"}
{"id": "train_46229", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "ci_integration", "tests_are_truth", "security_gates"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "00a8041795e2cac337bd80fa667df27b1a678134"}
{"id": "train_46230", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "evaluation_metrics", "tests_are_truth", "security_gates"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b658a01f35ab93c6756adda5bcfb5ae2bbf1f34f"}
{"id": "train_46231", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["governance", "reproducibility", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e7aa0c33a53b0fa105e5813f851aa0007a7d198e"}
{"id": "train_46232", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "security_gates", "governance", "tooling"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6cf2e1c12578e20d306aba5139d63cdd7119de36"}
{"id": "train_46233", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "repo_scale_reasoning", "security_gates", "ci_integration"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6577ac2b2fb6e979befc579b5b9b7159b7f3da61"}
{"id": "train_46234", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "tests_are_truth", "tooling", "auditability"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "96da1556bdc4f2a234c4a1c2d0beb5595c71112c"}
{"id": "train_46235", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "security_gates", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f104700d81732069f945b8a9e01f460e8818b609"}
{"id": "train_46236", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "evaluation_metrics", "reproducibility", "ci_integration"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "de1416fb3aaecb5f3e631a30c513e2f45bfbafd1"}
{"id": "train_46237", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "ci_integration", "evaluation_metrics", "security_gates"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9a9b236a7f47ce78277b375f8643fc5add5956f4"}
{"id": "train_46238", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "governance", "repo_scale_reasoning", "tooling"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "77fc10250ec5a135cef29768a6fbad4658e031f4"}
{"id": "train_46239", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["governance", "tests_are_truth", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d2ad0b7e57391a5a4b42a912fc61563ce3aeecac"}
{"id": "train_46240", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "auditability", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d36d17606321f4b601195ffe150df8001d382da0"}
{"id": "train_46241", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "tests_are_truth", "tooling", "governance"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7be01492e1356b2e941a3af6552916ab855a86b9"}
{"id": "train_46242", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "tooling", "documentation", "evaluation_metrics"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bea23bd133244b16b8899bb89e01b7b02cc84366"}
{"id": "train_46243", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "documentation", "repo_scale_reasoning", "tooling"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9d17620bd6b26264d245333eac2c78541f64c400"}
{"id": "train_46244", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "ci_integration", "governance", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9d5b19930f3c6fd9d6a6b0f283e8e96567d081a9"}
{"id": "train_46245", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "evaluation_metrics", "auditability", "documentation"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "16272ec565096873c55fb3bfa0f2a78cc4d34ccb"}
{"id": "train_46246", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "reproducibility", "auditability", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cfb5e82064d9c11b1be22d575debf2cb7cf8a61c"}
{"id": "train_46247", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "reproducibility", "ci_integration", "documentation"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "013aa2d425fe150eaee87392003867553f97356e"}
{"id": "train_46248", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "security_gates", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ea365c375ca654d01bc28d6d58bcca426089005e"}
{"id": "train_46249", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "evaluation_metrics", "tooling", "security_gates"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "600f35a9d938a785358459ae24839e935936ecc8"}
{"id": "train_46250", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "tooling", "security_gates", "governance"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cf29c4ccc637e23d65bd0cc0a1b5fe7ba9f1cb64"}
{"id": "train_46251", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "documentation", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a0b4f921a6ea118f115b2e795e874a3a47921d0c"}
{"id": "train_46252", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "evaluation_metrics", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a5c39deaecc079508783c49808fd91fd28624975"}
{"id": "train_46253", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "tooling", "auditability"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "08d21b3e5456356cf04fd4e3527700eea5d55af3"}
{"id": "train_46254", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "security_gates", "ci_integration", "auditability"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "250f06c6bf91e9f25e0d926d9f40b82fe57a74fd"}
{"id": "train_46255", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "security_gates", "tests_are_truth"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4be4f33ff40b8faa16bbcbb218cb3f117a5698fa"}
{"id": "train_46256", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "security_gates", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a75f131a765458de23d2faf35884085ea4dfa57b"}
{"id": "train_46257", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "repo_scale_reasoning", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "731a295a8a178fe203abb59800e92c3bdaa43174"}
{"id": "train_46258", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["governance", "evaluation_metrics", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6148203ce3af2e729317175744ecd3726eb664f7"}
{"id": "train_46259", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "governance", "tooling", "reproducibility"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e64f74928db65a124cf00640ea9ee30b46f1df3b"}
{"id": "train_46260", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "evaluation_metrics", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "da93adc0e65c1328dda5803e7e453171d4b9e468"}
{"id": "train_46261", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "documentation", "tests_are_truth", "reproducibility"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ae5877110e6c9df710ae022e1a460b797ed607f8"}
{"id": "train_46262", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "reproducibility", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5085baaeb6cffde71d0bb39d2c7bf36d8f8f9948"}
{"id": "train_46263", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "tooling", "reproducibility"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "40b04e9798db879e7aec7933c1dc09fc712b4b6a"}
{"id": "train_46264", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "governance", "tooling", "security_gates"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c0aae23422740a5ce0bfd505a95aa7ccf2e5f6b9"}
{"id": "train_46265", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "security_gates", "tooling", "auditability"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c0144fc461c195804c603cc8811b0d21fcf1fe3d"}
{"id": "train_46266", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "tooling", "documentation", "tests_are_truth"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "98afd21e4da187e5eba8ecf280279a403274d38a"}
{"id": "train_46267", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "auditability", "reproducibility"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "40d3fa3909133a961c3caeea0ee28af14019e672"}
{"id": "train_46268", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "ci_integration", "tooling", "governance"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6eb7be461c02fd3688039c8c8ae5b9740a8ecdf2"}
{"id": "train_46269", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "evaluation_metrics", "reproducibility", "security_gates"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "34f1400640ed7bff47fa0bed98f468a6d9d1964c"}
{"id": "train_46270", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "tooling", "governance", "ci_integration"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fcf8a20f9ce11900c02f2fd606cf85fe0ae6eaab"}
{"id": "train_46271", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "tooling", "evaluation_metrics", "auditability"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "02b2ee6a73873f6079cf1c50af66efbbca72c3fe"}
{"id": "train_46272", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "auditability", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0d4cb85cabe58bd8c179b0c45c4e91e9914086c0"}
{"id": "train_46273", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "tests_are_truth", "auditability"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "33ae5fb881442bf95cc381258153e2379128014d"}
{"id": "train_46274", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "documentation", "ci_integration", "tests_are_truth"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dd6c4a5f3c497ca129c904060ea61a7ef9a5d9ea"}
{"id": "train_46275", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "documentation", "tests_are_truth", "governance"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6219950162418951b7aa8033705cbf8855baf72e"}
{"id": "train_46276", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "ci_integration", "documentation"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1399b2101ee1ebada92e5a99a23acd83b4c746ea"}
{"id": "train_46277", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "reproducibility", "documentation"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4e57d2f2b459c32d7f2994a3c12021c7421490a2"}
{"id": "train_46278", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "tests_are_truth", "tooling", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8aa72b8db1d2abca8e1db8e2203bfc0801f15ebf"}
{"id": "train_46279", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "ci_integration", "tests_are_truth"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "515a842f56eee0769575e9351b201dd7a8110ea3"}
{"id": "train_46280", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "tooling", "documentation", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "39f85ead6ff16bb3772f87d3c3d66c2a3f7b5fe9"}
{"id": "train_46281", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "tooling", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "18ec966f1b2ed31e86bb49ac56dd92101a05ef12"}
{"id": "train_46282", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "documentation", "repo_scale_reasoning", "governance"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fd2855ea41c3d4c22b2348d0a39e3bb13f2b4242"}
{"id": "train_46283", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "tests_are_truth", "repo_scale_reasoning", "documentation"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cf4b210aec8b18fe47c086888b42808caa39b818"}
{"id": "train_46284", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "auditability", "reproducibility", "tests_are_truth"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "728970725c8e28ad64b3289a8fbea63d26c66b51"}
{"id": "train_46285", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "ci_integration", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cbc14cd93f18b2e2034f15c7a301baf4f2864808"}
{"id": "train_46286", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a5fdbed797bc051134daaf3254bcef9e137a2731"}
{"id": "train_46287", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "security_gates", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a9bfcd7675aad532cc4175b5a791fd88cdd83028"}
{"id": "train_46288", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "security_gates", "tests_are_truth", "documentation"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "198dd560bda3e99b971716f20c5e4eaef7a406d0"}
{"id": "train_46289", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "documentation", "tooling", "ci_integration"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "873c174b40a5cfe5bfc0c930c8dde9a1c6dbb405"}
{"id": "train_46290", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "documentation", "security_gates", "repo_scale_reasoning"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3a654abbc531f0c2bf5f4117d72b38fd5b5bbb66"}
{"id": "train_46291", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "repo_scale_reasoning", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6a9a21e3c98faa030eacbf1104e13cad187b2e26"}
{"id": "train_46292", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "tests_are_truth", "documentation", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7cd81a5b60b59ddc08cc99e2cd9778e8a9ea8442"}
{"id": "train_46293", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "governance", "auditability", "ci_integration"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "78e7debc9fc71d98859ac74ca7daac9f56dedec3"}
{"id": "train_46294", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "ci_integration", "documentation", "tests_are_truth"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4e240484c94609a8bfba9abf411c01eb263c0271"}
{"id": "train_46295", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "tests_are_truth", "governance", "documentation"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "31717dde7d7dbcdc0c2d2d6013c4fbfab9159d83"}
{"id": "train_46296", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "cost_latency_tradeoffs", "reproducibility", "documentation"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b4e5510362efcb7b042b42d1db4818e0745c345a"}
{"id": "train_46297", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "governance", "auditability", "security_gates"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4510cefbcded174946a44533b33e21635eb275d3"}
{"id": "train_46298", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "tooling", "evaluation_metrics", "security_gates"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "27b64f9d96c50a81e7ac05fa0bb9b8e5810e3545"}
{"id": "train_46299", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "evaluation_metrics", "security_gates", "tooling"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "72bfb6f57bf2df0701e45b84e32a4f61f227348f"}
{"id": "train_46300", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "evaluation_metrics", "security_gates", "governance"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b80bf8c1f19e711f875bec8f7b379fd1e72fe5b1"}
{"id": "train_46301", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "governance", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "052ef3ec5936d5b5a2e97cd43d7c2ae4cc363905"}
{"id": "train_46302", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "tooling", "reproducibility", "evaluation_metrics"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e43d81de8cd280383fc7abe019771b15ccbdd3c9"}
{"id": "train_46303", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "security_gates", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "297f7039eb6a54e782f136e9d21c2edeefa6065b"}
{"id": "train_46304", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "tooling", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c4a1e0c127928022ee0731cab3b03a3e9637872e"}
{"id": "train_46305", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "tests_are_truth", "evaluation_metrics", "security_gates"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ada1fa5191467ca19a10c80d8d9ae367623678da"}
{"id": "train_46306", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3af546649f9ea69b8e93dcda2f4e4b7023e39911"}
{"id": "train_46307", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "ci_integration", "cost_latency_tradeoffs", "governance"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a0c3df38e8d3ffddfb9c4d5a0cadfb9ad5530a03"}
{"id": "train_46308", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "tooling", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "62c7371b96b1848ed5a1ec428a499d98500b03c4"}
{"id": "train_46309", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "ci_integration", "repo_scale_reasoning", "auditability"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1a713524317b97115197d8f1e09946ed95969bfc"}
{"id": "train_46310", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "documentation", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ccfb3cf813de1f1798fdc5a890219a7fd0c498b9"}
{"id": "train_46311", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0d3f491449d7093d4bbb3a8abe4fe59b4b2fc945"}
{"id": "train_46312", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "security_gates", "reproducibility"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "587c3810804546943e00c663dfb81ac14947cb9c"}
{"id": "train_46313", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "evaluation_metrics", "ci_integration", "tests_are_truth"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "731a295a8a178fe203abb59800e92c3bdaa43174"}
{"id": "train_46314", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "reproducibility", "documentation", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5523c44c446e069fe3129a5906f1979879cf014f"}
{"id": "train_46315", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "tests_are_truth", "security_gates", "documentation"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ea0506b9179b2beb65facba33f9904a71ac892b2"}
{"id": "train_46316", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["governance", "reproducibility", "repo_scale_reasoning", "auditability"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "23847aec4837cdae28d64bab74fe4ae4ec526569"}
{"id": "train_46317", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "tests_are_truth", "governance"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "40f10eb640912e482b355a91876e841af07e7a71"}
{"id": "train_46318", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "evaluation_metrics", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "54ae2ded17b52986ece0c0ee9d07ecfb3eedabe6"}
{"id": "train_46319", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "tooling", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8682f66a993826539d970748fbeba782ec900921"}
{"id": "train_46320", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "tooling", "security_gates", "evaluation_metrics"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dc121c088a9fceeadf87af78c4ab3dca68bd0a40"}
{"id": "train_46321", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "documentation", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ef743eb10e5038d33f26555e2ee1aa0cb249b2b1"}
{"id": "train_46322", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "evaluation_metrics", "ci_integration", "documentation"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "181e29d9e387c90224fceb55c4a203373528ecef"}
{"id": "train_46323", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "security_gates", "cost_latency_tradeoffs", "governance"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0004acf31fb01a12ecd583dd4b4c7343c895d495"}
{"id": "train_46324", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "evaluation_metrics", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7e50de9444c0e304dedf40739be1ed74bae7e226"}
{"id": "train_46325", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "documentation", "ci_integration", "security_gates"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5cd7c21a06c594ace42833d70c7d273ddd049fbf"}
{"id": "train_46326", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "tooling", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9d4d214bd96dc715b26e1f4c7c98cbc69e8cd667"}
{"id": "train_46327", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "documentation", "governance", "repo_scale_reasoning"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "55ded0c70438c26998c4e50c44eb1c2970c0590f"}
{"id": "train_46328", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "auditability", "reproducibility", "documentation"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "461283839b6a1260d611379c35794a9d9533f543"}
{"id": "train_46329", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "security_gates", "documentation", "tooling"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d0cb83836b70a249b4581176e3868d405ef755b5"}
{"id": "train_46330", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "evaluation_metrics", "tests_are_truth", "tooling"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d359f449eb60183ba2b943a12a0d4c49e68e3d67"}
{"id": "train_46331", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "tests_are_truth", "documentation", "auditability"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "855fd810ffe59c02b166509d0e2abd6d5aad6062"}
{"id": "train_46332", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "ci_integration", "auditability", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e18f2807c09e8c3139075a9d8f7b50521ac772bc"}
{"id": "train_46333", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "reproducibility", "governance", "evaluation_metrics"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a58271d0d5750b249dc03779502341f440e81e9f"}
{"id": "train_46334", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "reproducibility", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f709037f1d47909678330247397243f4bf917293"}
{"id": "train_46335", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "tests_are_truth", "security_gates", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "12055db54c11db8da6b9ecdb03cb4193fc36825c"}
{"id": "train_46336", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "documentation", "ci_integration", "security_gates"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ea9c267af746c0515192a86ebf69157824a3c260"}
{"id": "train_46337", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "repo_scale_reasoning", "tests_are_truth", "tooling"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c78b5b29cd8363699571e2899cf26980f7a44bef"}
{"id": "train_46338", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "auditability", "ci_integration"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a925f722d915af0d77f0eeaeb17a5df1b73ddbfb"}
{"id": "train_46339", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "security_gates", "tests_are_truth", "ci_integration"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8ca506648435c306476a7c84d0aa81a2c568ebe6"}
{"id": "train_46340", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "tooling", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "942a7a28460223375d6da5bbce69bd29c8749c49"}
{"id": "train_46341", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "evaluation_metrics", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ab21195765124b88c88de0a966ab6aa2071ef63d"}
{"id": "train_46342", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "tooling", "repo_scale_reasoning", "governance"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "baa9b17a2c96e5d923fe24c755f41ca01f880f58"}
{"id": "train_46343", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "tooling", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ee85eabcee6e4074a97bfb4395155cdbb0c57c65"}
{"id": "train_46344", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "documentation", "security_gates"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "978ab70870cefbbc392980ab052f9c12af64d592"}
{"id": "train_46345", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "reproducibility", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1c455ab04c617100aef869aa5bb7a6aad315cc80"}
{"id": "train_46346", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "evaluation_metrics", "auditability", "ci_integration"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0cfb7c1013f739e690bf6a2d7f195f24a39d7570"}
{"id": "train_46347", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "governance", "ci_integration", "evaluation_metrics"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e8d9cb1da200d672fa508e3df6ff66beba9f8968"}
{"id": "train_46348", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "tooling", "reproducibility"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "947a3111adccfae2ce5078b1d28b54c679b20586"}
{"id": "train_46349", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "reproducibility", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d3cc84c0a7141bdcc758b2c29322e923a93e56f8"}
{"id": "train_46350", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "auditability", "reproducibility", "evaluation_metrics"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "57423224c023b38d94062e7dd0c5e0d09e396327"}
{"id": "train_46351", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "tooling", "security_gates"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6a81351edb899ed6fd02359f2d344f2b57fdffbf"}
{"id": "train_46352", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "tooling", "documentation"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "589cea79f06b30bf029e112cb4d7d27d77c646c4"}
{"id": "train_46353", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "governance", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "59b865ba76dbb2ae0a891fa7ad1da3c76a7ead78"}
{"id": "train_46354", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "cost_latency_tradeoffs", "tests_are_truth", "security_gates"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fad8c3298a21820a649afa41e011bfbfda5741f8"}
{"id": "train_46355", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "security_gates", "governance", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cfee27f84384818fa42974d59c29f7de25ebbd64"}
{"id": "train_46356", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "ci_integration", "cost_latency_tradeoffs", "governance"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bb6f9348ee40abe077a5926a338613bdee6d4f5a"}
{"id": "train_46357", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "ci_integration", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4379adf174aaa657dc98fcf734d986dccdef72da"}
{"id": "train_46358", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "reproducibility", "documentation", "auditability"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "46f2f02a81d5b6b77d9b1bf48c809a1b49d5afee"}
{"id": "train_46359", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "documentation", "reproducibility"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "febb6c86f1d15ccd05990cacf9438e06006fdd44"}
{"id": "train_46360", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "repo_scale_reasoning", "evaluation_metrics", "tooling"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f056bc43efe55d233a65e4aa2b0544bd5c4aec90"}
{"id": "train_46361", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "documentation", "governance", "reproducibility"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7428259bbfd54a068ecc5ab6b5deb1cec3c43027"}
{"id": "train_46362", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "tooling", "repo_scale_reasoning", "security_gates"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "05ec1c562f852a1a620b3f115a39ed8d6d17fc15"}
{"id": "train_46363", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "auditability", "ci_integration", "tooling"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e58b2c47d19dccf02ebe98992994c44f342121f3"}
{"id": "train_46364", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "tests_are_truth", "reproducibility", "auditability"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "83cdd776448499657509bc3c9fe6f560c62bd2be"}
{"id": "train_46365", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "repo_scale_reasoning", "documentation"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8dcf8c926bc8d62b48140cfc47c131c2f6083253"}
{"id": "train_46366", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "auditability", "tooling", "documentation"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5d257b21a6ea355e1399874eba0c29fa0278de05"}
{"id": "train_46367", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "tooling", "security_gates", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bfddcc442b1eb89ee450677b68e59c0fff813e36"}
{"id": "train_46368", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "tooling", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e832c9f95ab85484a80cb3852f946debfee2f631"}
{"id": "train_46369", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "58fd9e4b99dbb84dd3226443cf549e2263a1c4a7"}
{"id": "train_46370", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "reproducibility", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cc6c5adf1df28ae0b5dc363fe7eea8187a2e46e2"}
{"id": "train_46371", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "reproducibility", "repo_scale_reasoning", "documentation"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a969d64959ca233183dcd9baddaaa832b62f0739"}
{"id": "train_46372", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "security_gates", "reproducibility", "ci_integration"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5541c4ff67202a9a2b6839c1f72211fa81f8fb6a"}
{"id": "train_46373", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "tests_are_truth", "evaluation_metrics", "ci_integration"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9a30c4d0ec13572439ca33bf9bdc0bd24e6fe724"}
{"id": "train_46374", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "cost_latency_tradeoffs", "repo_scale_reasoning", "security_gates"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a698f76adb2f48cf25f995b007fa6e93e63ad671"}
{"id": "train_46375", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "governance", "documentation", "repo_scale_reasoning"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9db365cad2e0283b4e72d8ab9222e1fe9e37500d"}
{"id": "train_46376", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "tooling", "governance", "reproducibility"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1a586fd0b613ee84f9fa119a406f8033bd8c2f32"}
{"id": "train_46377", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "security_gates", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f2dc2ebef4f545f2420338455f4d257999ddae66"}
{"id": "train_46378", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "auditability", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "102a6642efa770721e97bc6bb65ccc7c789630cd"}
{"id": "train_46379", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "ci_integration", "evaluation_metrics", "reproducibility"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "93a63a0cc2e5140e9b5997253ccb26877cb61f0a"}
{"id": "train_46380", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "auditability", "tooling"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "622d6981b9650e6ebca804120c064a1b42374889"}
{"id": "train_46381", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "repo_scale_reasoning", "evaluation_metrics", "security_gates"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c87fd097567db7d6fc3f4bb9d5f1c43ac99b3af6"}
{"id": "train_46382", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "documentation", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d5d3ad2c54528e9294f37fd0ed1d596f272896fe"}
{"id": "train_46383", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "documentation", "tooling", "reproducibility"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "41581516a503c00b994fc151f71dcd1b04dc20f2"}
{"id": "train_46384", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "ci_integration", "auditability", "tooling"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dd3dba469a31412ff0f31a35f351587d41b2ce2c"}
{"id": "train_46385", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "governance", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dc02682ccd83cfbf97f47121bdde6258bdad6ced"}
{"id": "train_46386", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "evaluation_metrics", "reproducibility", "ci_integration"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4b9efda42661b6fda9df58abddef68e7577686b2"}
{"id": "train_46387", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "tests_are_truth", "tooling", "security_gates"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cd8eaa008f85d0571a76f302180f558fd137c88a"}
{"id": "train_46388", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "reproducibility", "security_gates"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f07716b4d877bb05bc44bc043968dfaf61621602"}
{"id": "train_46389", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "documentation", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b2fff2356287a1d9dae7f8e8df91bdfe60d37322"}
{"id": "train_46390", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "documentation", "governance", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4052750e833a358f2137d142f339c76ffaf8f560"}
{"id": "train_46391", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "governance", "ci_integration", "evaluation_metrics"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b335e901c2e5d0885d12d50ab5854f61b2202bdb"}
{"id": "train_46392", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "ci_integration", "governance", "reproducibility"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "999877a4824868e2c2d772eac74410ae8d282798"}
{"id": "train_46393", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "documentation", "reproducibility", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "46edf4dca9073c7bbcdc47c82aaea37f56484d03"}
{"id": "train_46394", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "ci_integration", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "519a24d8c15af9f7882baad413173c792b10394a"}
{"id": "train_46395", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "evaluation_metrics", "reproducibility", "security_gates"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fc62b09014f22a12a3a2bb5343f72c4f468482b0"}
{"id": "train_46396", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "documentation", "auditability", "tests_are_truth"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f7b8776c201bc687e4ca0f922250f3c5bd7b1e50"}
{"id": "train_46397", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "documentation", "security_gates", "tooling"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d9987d3d90900bf525ba8a193133a3bbc0f866bb"}
{"id": "train_46398", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "security_gates", "evaluation_metrics", "documentation"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6ffdc5ff26d73bcc35c4f21268bbc93b96db0eaf"}
{"id": "train_46399", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "governance", "tests_are_truth", "ci_integration"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "78de76474794f5cbfc0607923d185f6e6b6ca125"}
{"id": "train_46400", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "evaluation_metrics", "documentation", "tooling"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c4e76ee168185313eb639c94ae45ef50dfafa798"}
{"id": "train_46401", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "documentation", "tooling", "security_gates"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ed1aa7f501d57db81279566c861d697023bf3cbd"}
{"id": "train_46402", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "ci_integration", "security_gates", "tooling"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "23f150093184637cf60f4de287b50e8501c4f051"}
{"id": "train_46403", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "documentation", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "196ef74efb445ba7d1b526b7328e0f5e82ab99ed"}
{"id": "train_46404", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "governance", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7203cc3cef03f179953ce5762232fac758da566c"}
{"id": "train_46405", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "ci_integration", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "941c2aac652783ca05ae1e8de5d437513532cf4e"}
{"id": "train_46406", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "auditability", "reproducibility", "tooling"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "31b2bae43143e686aea1c2ca0409d3723e103b87"}
{"id": "train_46407", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "evaluation_metrics", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fc407cfdb23e61f29468f0192a2397b01059dd95"}
{"id": "train_46408", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "documentation", "tests_are_truth", "reproducibility"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "50f449d6f9eb41588c6afeca595d36e98735a6ee"}
{"id": "train_46409", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "tests_are_truth", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "648230042ff698ded340c930140a9d7025774b6f"}
{"id": "train_46410", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "reproducibility", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f2dcc32d3defbe8d98f118d74360172068ba61b1"}
{"id": "train_46411", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "governance", "auditability"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ff9011851087891713931bb00b9ab426b96422b1"}
{"id": "train_46412", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "evaluation_metrics", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "262b351f86dab7ef01a823c62f9958a97f479b28"}
{"id": "train_46413", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "evaluation_metrics", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1c165027dae0f0b7a8ee27bb379cb49cb9ebe132"}
{"id": "train_46414", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "reproducibility", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1a1c47f5942f6bc61f2b1d1b4943d1aea62a46ff"}
{"id": "train_46415", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "tooling", "security_gates", "ci_integration"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "704bb005731101e7369696ba87f47553af503f7e"}
{"id": "train_46416", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "reproducibility", "auditability", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e2902daa1b4dcc91436e83a8fbb3afb5e9de4866"}
{"id": "train_46417", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "governance", "tooling"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0d8fe929e9d185977c6671a8c8687f5313e0a369"}
{"id": "train_46418", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "reproducibility", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ee11923a4737b2c062dd2c1ef26a3a8864e2ee0b"}
{"id": "train_46419", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "reproducibility", "ci_integration", "tests_are_truth"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fd6a9259fb897eddceb362d34f6efcedcbd0646d"}
{"id": "train_46420", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "tests_are_truth", "ci_integration", "documentation"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6a712f43dd1a65061cddd123a2865769f40b8346"}
{"id": "train_46421", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "evaluation_metrics", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0d7b8335485eaf4a36ad53df86a3253aaa624794"}
{"id": "train_46422", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "documentation", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3b8ca3694f100ec11ad7ab964f74cc3b94fd30d8"}
{"id": "train_46423", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "reproducibility", "tooling"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "738e59870d9cec94390676c2fc93e1a423b97622"}
{"id": "train_46424", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "governance", "evaluation_metrics"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "028587454e1e70fc03f16ebd05c31f831bede196"}
{"id": "train_46425", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "security_gates", "reproducibility", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "42ad6bba4f2414071248bae9c8a0e8a3ee64b180"}
{"id": "train_46426", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "evaluation_metrics", "documentation", "ci_integration"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f911eb2fb55cb04c3c7caf7af5ba9fa031d2d3ff"}
{"id": "train_46427", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "ci_integration", "security_gates", "tests_are_truth"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4c75496836e82413bbb0edae9888a6f99c499ba3"}
{"id": "train_46428", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "tests_are_truth", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0c57f739f3c0a7c357e43b0afc06b9831f8e2306"}
{"id": "train_46429", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "repo_scale_reasoning", "auditability", "ci_integration"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e73f1618ed7561e8d53af404bf4785dfdb4db797"}
{"id": "train_46430", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "evaluation_metrics", "ci_integration", "reproducibility"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "61066ebeeaf48860a9b36159f1a8329be22e0d21"}
{"id": "train_46431", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "governance", "auditability", "tests_are_truth"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1f8a1dbbaba54dddbad007bb2466ee34536c6d2e"}
{"id": "train_46432", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "tests_are_truth", "evaluation_metrics", "security_gates"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b4ab948815c9d6e152a8f15319849ea126035e9d"}
{"id": "train_46433", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "auditability", "reproducibility", "tooling"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "41b92495c66e5021d19fa6fb15f9a226017bec42"}
{"id": "train_46434", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "documentation", "governance", "tests_are_truth"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b2cacd75e20657e429ee4dec3bb6638ba7a5bbe2"}
{"id": "train_46435", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2d079af68b948c24f7d283cb3e5bab4e2946d24e"}
{"id": "train_46436", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "security_gates", "tests_are_truth", "ci_integration"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c6e65d8391f361c5556d501bbfe96dd42c763750"}
{"id": "train_46437", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "documentation", "governance"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c78897e1f0014015e0213b46182bd7ca759bd7f3"}
{"id": "train_46438", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "tests_are_truth", "governance", "auditability"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1792a90830716f1873cf0297874af708ce6d93de"}
{"id": "train_46439", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "governance", "auditability"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ee280604651a9d250d4c6249396a957814215a59"}
{"id": "train_46440", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "governance", "tests_are_truth", "reproducibility"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6d84587e58f206bb969094dce3afed5bd5acdfe9"}
{"id": "train_46441", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "security_gates", "evaluation_metrics", "tooling"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "edbb0e1ef7481088d67c065d8557e2f11b7b249e"}
{"id": "train_46442", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "tooling", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "92a6b00788a1988bd778236ff39a45ad5ec52557"}
{"id": "train_46443", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "documentation", "repo_scale_reasoning", "security_gates"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e678393b446d6979b1dce6f3a05f7051f5db3cb6"}
{"id": "train_46444", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["governance", "tooling", "security_gates", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "35c3f258213264d0fb9a8e22d9e1ab5f0efc419e"}
{"id": "train_46445", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "documentation", "ci_integration", "evaluation_metrics"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cfae7e09df4e92acf2f6c1b6e1a034ad7514fc25"}
{"id": "train_46446", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "security_gates", "tests_are_truth"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c98c1443db7fdaae2956f01687510bd09354de42"}
{"id": "train_46447", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "auditability", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "68085678c5dc3be9b15d5276a28963e3c79f67c2"}
{"id": "train_46448", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "auditability", "ci_integration", "tests_are_truth"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "73cf9767179918833811b6390e15a92ac85fcd26"}
{"id": "train_46449", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "evaluation_metrics", "repo_scale_reasoning", "governance"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5ea155024e26b1c85c7b4e35bd7459575c2735b7"}
{"id": "train_46450", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "documentation", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4c6625fc857f18e99e1d6c178acd7311d0f4d8ca"}
{"id": "train_46451", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "governance", "auditability", "tests_are_truth"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "270bb1d3976f9e31589497e5dce1dd8751a82818"}
{"id": "train_46452", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "documentation", "repo_scale_reasoning"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bb411324b0d265e5780e540c79c8ff41f0cc6ec9"}
{"id": "train_46453", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "governance", "reproducibility", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "77d4ff7be59b38e1d883411227fc1f19cd574ed1"}
{"id": "train_46454", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "reproducibility", "governance", "documentation"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cd393c150e7d04de8d4318787d478aee5fc068d3"}
{"id": "train_46455", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "ci_integration", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ba3299fdfa840d59556ea10ecf02bcc0345f0625"}
{"id": "train_46456", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "tooling", "governance", "auditability"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "66d48c0f064a6f81ffd4e1896acaba4db619469a"}
{"id": "train_46457", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "security_gates", "auditability", "evaluation_metrics"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7b97a24b4266586e8b5d032c89485350fa332783"}
{"id": "train_46458", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "governance", "security_gates", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "661de9065f792bcf0509468650f163f62eba53b7"}
{"id": "train_46459", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "auditability", "documentation", "security_gates"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aedcc472a66373e528806de6810cce5b74d77965"}
{"id": "train_46460", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "governance", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ed1c7c0d8e9a60a59114aeaa7097e6fce4d8ccb1"}
{"id": "train_46461", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "documentation", "security_gates", "ci_integration"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6af4c52dd12b9f64e34b5e67599b2c8906879791"}
{"id": "train_46462", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "evaluation_metrics", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "689f6e288216779816983d7ed9a291783ba35d7b"}
{"id": "train_46463", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "tests_are_truth", "reproducibility", "ci_integration"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9fe8bd4c0b2fd3a82dc1f1a2d0bbc59b3e321222"}
{"id": "train_46464", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "auditability", "reproducibility"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5aba9c280f2a3fad0d7266b70a5eac977d5b0004"}
{"id": "train_46465", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "evaluation_metrics", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4a4e204dab9c46764c20349862d20a119d5ef347"}
{"id": "train_46466", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "auditability", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f9b47b9af42f23688097c83c8cee5831480e3d55"}
{"id": "train_46467", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "governance", "repo_scale_reasoning", "auditability"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b264c661a517c1635a375d608ef0b61e8e56ef4f"}
{"id": "train_46468", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "documentation", "reproducibility", "auditability"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "19dd3677ab80704a8b7ea56a8c56c7bbbaa435ef"}
{"id": "train_46469", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "evaluation_metrics", "documentation"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2e8332d245b1dd046fb9333dd6106ce099b6e087"}
{"id": "train_46470", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "governance", "auditability"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "28a03a9fbb18acb3b0aace8eaeda3ddbe1bfb186"}
{"id": "train_46471", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["governance", "tooling", "security_gates", "auditability"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3fab3f21cf7eac765b84bdcfa3fbc934032729cb"}
{"id": "train_46472", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "cost_latency_tradeoffs", "governance", "reproducibility"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "25f983f0e5a4093b86735811f60ae0e7484c422d"}
{"id": "train_46473", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "security_gates", "tooling", "auditability"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ea11daaec00571dd191b656a1b4555ce1782d2f8"}
{"id": "train_46474", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "tests_are_truth", "tooling", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dc69c5fd5cfe84a83c134ee18fea0975846f0b07"}
{"id": "train_46475", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "auditability", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e58b2c47d19dccf02ebe98992994c44f342121f3"}
{"id": "train_46476", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "cost_latency_tradeoffs", "security_gates", "governance"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b994cedd70e0b935ae58528ee4a141014f88309b"}
{"id": "train_46477", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2a107e456f1c10908be529b2b4a68c8ca85432a2"}
{"id": "train_46478", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "auditability", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e9492808c8825ff19f183cfffa2af785b37185ec"}
{"id": "train_46479", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "documentation", "governance", "ci_integration"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4ac8a8df5aa8659606bcd94700d25400c1a290f2"}
{"id": "train_46480", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "security_gates", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "144ed43f4dfb22253303006022df3144c5c05a77"}
{"id": "train_46481", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "tooling", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1323900fd0d2d5ab37ab2b80eff46d315d6ded76"}
{"id": "train_46482", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "reproducibility", "tooling", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e3fb5189fe3ccf266b0a665423cac84eab509797"}
{"id": "train_46483", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "tooling", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "94bd31124501164e37273230d8a45851f6b60c46"}
{"id": "train_46484", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "security_gates", "evaluation_metrics", "documentation"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2c36e46e8398d28df87da9343d8964bfb9090158"}
{"id": "train_46485", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "governance", "tooling", "evaluation_metrics"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c1d1e10ef17e57ae88f7297b3cbcbf8cac7d35bb"}
{"id": "train_46486", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "reproducibility", "ci_integration"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "255b185fc9f3720c7e950f719c173740a940ce70"}
{"id": "train_46487", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "security_gates", "documentation", "tooling"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "016a8df867549d7005d299c033ea5409b45c8de6"}
{"id": "train_46488", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "repo_scale_reasoning", "tests_are_truth", "governance"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e3bd48be3ec349ab7ccdd68a532bbd2aa75b9bcd"}
{"id": "train_46489", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "documentation", "tooling", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fcd7a6a2991d14aa21e5bd3fd430a6636447dd7a"}
{"id": "train_46490", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "cost_latency_tradeoffs", "reproducibility", "security_gates"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9e4103bfc0585fb48f26c1e95ddefd37129b6b6c"}
{"id": "train_46491", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "evaluation_metrics", "tooling", "governance"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "250d6e03e2f97a7f18b115dacfe54e93b6420cce"}
{"id": "train_46492", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "reproducibility", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e2c8184320c871ba452fcda115e328ea942bef36"}
{"id": "train_46493", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "documentation", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "215b0a8e0393adbbda41529b4f09ce711862686a"}
{"id": "train_46494", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "governance", "security_gates", "auditability"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "62b6a5e234c205b287efbd58a80390f4373d7bf9"}
{"id": "train_46495", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "repo_scale_reasoning", "ci_integration", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2c61d2f6b64992bb03c322b3dd618818164ed25c"}
{"id": "train_46496", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "tooling", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "001df8d625420b32aaba69ce9b2248774ebe543d"}
{"id": "train_46497", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "security_gates", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dc60306c05c2612a794766eaf7356a95ea7717a1"}
{"id": "train_46498", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "auditability", "tests_are_truth", "governance"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d7076693ac4530d82345cdfece7cd2c7ca7965c3"}
{"id": "train_46499", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "tooling", "reproducibility"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f1cd3297e9de2e0a13bb0a2a4d93e14b5190e4ae"}
{"id": "train_46500", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "evaluation_metrics", "governance"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1da9e360894aa68bf3ed84177a7304cb84af4e5e"}
{"id": "train_46501", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "documentation", "reproducibility"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5032ef50ba56b1f10458f3cfad34f49964ce708d"}
{"id": "train_46502", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "governance", "security_gates", "tooling"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2cb223fb4678fc3e278d5e0ac658038da5d66aec"}
{"id": "train_46503", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "tooling", "auditability", "ci_integration"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "221d802af38eb1cd86fce9f34f66275c22c32a80"}
{"id": "train_46504", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "auditability", "documentation", "tooling"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0592571c7cd1591ab65fc8b4c4ef74be48d4df51"}
{"id": "train_46505", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "cost_latency_tradeoffs", "ci_integration", "documentation"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a73388f59d83b37e1409b234ace6729783d3ff3c"}
{"id": "train_46506", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "evaluation_metrics", "documentation", "reproducibility"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2a71e0774cd394a18455de56ea0e60ceb9f60145"}
{"id": "train_46507", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "governance", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "13debddf7ff43169c5bcf5effed7b1b7e9ce1151"}
{"id": "train_46508", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "cost_latency_tradeoffs", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0e8338a247da74c90fb31648746a2f551bccbf25"}
{"id": "train_46509", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "evaluation_metrics", "security_gates", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "77fdf3afbfa346c36a16a881e9efee812bf1350f"}
{"id": "train_46510", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "auditability", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f42d9f9702aad676c4878230909df787ad1570cb"}
{"id": "train_46511", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "auditability", "documentation", "reproducibility"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0a7ba41331c87815d5814058e71327ab51aa3877"}
{"id": "train_46512", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "documentation", "tooling", "auditability"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "79f3dfe8aab2df1d277713dccc8d4049d959cd91"}
{"id": "train_46513", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "security_gates", "documentation", "evaluation_metrics"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc22701dcdcc1289818ccb0abac03a4de40df7c9"}
{"id": "train_46514", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "security_gates", "reproducibility"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f84af4a6996eeb4976689b9186fc2407c97e196f"}
{"id": "train_46515", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "repo_scale_reasoning", "tooling", "evaluation_metrics"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4165845815a95d1b3c8f060a6fc62d75fbad1d51"}
{"id": "train_46516", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "ci_integration", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2c67501a550541a3cc7737d8bbbb33a1f694c5ce"}
{"id": "train_46517", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "security_gates", "reproducibility", "ci_integration"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5048f0b37380a119b42fbfc5791e0d3b0dd186da"}
{"id": "train_46518", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "ci_integration", "reproducibility", "tooling"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "11e46ad6f2af649fac7a42bce22bba6ab882af9d"}
{"id": "train_46519", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "tooling", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "60c4fd2820193af11e6c4e26ff6c0137cd234f5c"}
{"id": "train_46520", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "ci_integration", "tooling", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8de0e11e91e4483ab3b1e2c5c92c30eb8d63e07b"}
{"id": "train_46521", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "auditability", "cost_latency_tradeoffs", "governance"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aed8eb0cfaa56de2938213e3cc554552c8666796"}
{"id": "train_46522", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "auditability", "reproducibility", "security_gates"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "00b8c37535032a7546c3f78118a34a80a926e8e7"}
{"id": "train_46523", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "security_gates", "ci_integration", "reproducibility"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1e5bd38833d20be11d3ab43730f884c444578bec"}
{"id": "train_46524", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "reproducibility", "security_gates", "ci_integration"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "813619a79aaf18ff0d10e521e5f19965c5b77e44"}
{"id": "train_46525", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "tests_are_truth", "security_gates", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6d84587e58f206bb969094dce3afed5bd5acdfe9"}
{"id": "train_46526", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["governance", "tooling", "evaluation_metrics", "auditability"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "22184ebe96d6a0b262273756135b9cfcb3e401e0"}
{"id": "train_46527", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "evaluation_metrics", "tests_are_truth", "governance"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "346da62fd47abd2fe23340414011d70973a2abc6"}
{"id": "train_46528", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "security_gates", "ci_integration", "documentation"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b29853a339741678a579dd988f9d20c5a5f46a59"}
{"id": "train_46529", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "security_gates", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1bcd784b9dba6574ae341de9e0a38b6bd87de2fb"}
{"id": "train_46530", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "cost_latency_tradeoffs", "tests_are_truth", "reproducibility"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "870940ef96d3a3d726323dd757ee08d744a20acd"}
{"id": "train_46531", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "tests_are_truth", "governance", "security_gates"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "343f2b97f8d5d6007e715c256347c0e15120fc07"}
{"id": "train_46532", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "repo_scale_reasoning", "reproducibility", "auditability"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "af407e6ef490c002db48f5bcdb63048d94436b46"}
{"id": "train_46533", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "tooling", "governance"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "06d48d3dd6ce549d9db6b89bb838a273b299f16f"}
{"id": "train_46534", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "repo_scale_reasoning", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fd3353fd2a27f5b60ef8a6074b00007c906c6530"}
{"id": "train_46535", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "auditability", "evaluation_metrics", "security_gates"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "700536162c451fac61b628077ff8237d6a4179b8"}
{"id": "train_46536", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "repo_scale_reasoning", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f46288e9a505f6766774f18988031329e49863e3"}
{"id": "train_46537", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "auditability", "governance", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e8ae01fdf451ac8c733813dcbd59429c8b1b0389"}
{"id": "train_46538", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "auditability", "security_gates", "reproducibility"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f443b4a6928906da8860c5f1d186d3d105477dcb"}
{"id": "train_46539", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "tooling", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "227154eab4121793b56be8b353688e615909de5e"}
{"id": "train_46540", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "ci_integration", "reproducibility", "tooling"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a38041335c201c52edd01f17d1e697edc7c48e10"}
{"id": "train_46541", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["governance", "repo_scale_reasoning", "tests_are_truth", "documentation"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c2789583a0b0bd35b30afced641088fb2c70258b"}
{"id": "train_46542", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "documentation", "security_gates", "reproducibility"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "416ee131a96f5a54be9fcc10ef06ffb8eeb8b8b7"}
{"id": "train_46543", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "tests_are_truth", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "60abcca4a316474de5e2a3a738a87091b19e92bd"}
{"id": "train_46544", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "tooling", "cost_latency_tradeoffs", "governance"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4ad02c4fc05ed72fbd52672717cce316db8fc96f"}
{"id": "train_46545", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "tests_are_truth", "security_gates", "documentation"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "004d08eac9cd12545e1a18ab629e5ccde8b3f081"}
{"id": "train_46546", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "evaluation_metrics", "reproducibility", "ci_integration"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "82baca7300b20610db92607e1488493a5d8d9c8a"}
{"id": "train_46547", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "repo_scale_reasoning", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3d7eb4eea714c2a866f28099e658b6ac04c854da"}
{"id": "train_46548", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "reproducibility", "governance", "auditability"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d055e06cb645861cd981b04e0f8c5480270ac1e1"}
{"id": "train_46549", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "documentation", "auditability", "tooling"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ef96c06b568f12dc72a6ca5eec4e81085e595f80"}
{"id": "train_46550", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "reproducibility", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "567f58a4aa5ca12052118d80b977dc2778fe1d4d"}
{"id": "train_46551", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "governance", "tests_are_truth", "tooling"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6577979ce00a30e50e550498f3465cadacbe4ff3"}
{"id": "train_46552", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "repo_scale_reasoning", "tooling", "evaluation_metrics"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ea4e116f17623d329d2fd6b42da6a746c6123006"}
{"id": "train_46553", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "auditability", "tooling", "ci_integration"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b4892d94505c78cac9d1440cef5d3b148e04084f"}
{"id": "train_46554", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "auditability", "tests_are_truth", "ci_integration"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "015642b4a80044615fb91a2fb77d56699b3112d2"}
{"id": "train_46555", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "evaluation_metrics", "auditability", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1deaf8938c0e1ab3455f84ea54367ffe232bb5b0"}
{"id": "train_46556", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "governance", "evaluation_metrics", "ci_integration"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fb27f496e101c0e14cdb9884b65f95ca2e489060"}
{"id": "train_46557", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "tooling", "documentation", "tests_are_truth"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a6ed3ddca2981feb0a13bfe7311ad3ef1ab40e8d"}
{"id": "train_46558", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "documentation", "security_gates", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7858744b5023807e551a9dd4dd539f421fdf80a8"}
{"id": "train_46559", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dcf9fd7cc2a550e5e3f0080899c8ab2076920ddb"}
{"id": "train_46560", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "ci_integration", "tests_are_truth"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b79206bc2111b6c8508cb7e6cc3c2f398b7296e1"}
{"id": "train_46561", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "documentation", "repo_scale_reasoning", "tooling"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c7a2a63f744ab62df956211ba9dbaa7d1d7ddb86"}
{"id": "train_46562", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "auditability", "security_gates", "reproducibility"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f5c9cd4adce404671eb9fda1159739effdf70182"}
{"id": "train_46563", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "tooling", "governance", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "63cbc72a41bed999d8f83ad24d72c51f8313144a"}
{"id": "train_46564", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "security_gates", "auditability", "repo_scale_reasoning"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "59fefff0062fec7532a5f191e8051c57e5cf900f"}
{"id": "train_46565", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "tests_are_truth", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1d4394b8629c48e056b1c78a8d3c4194daaaa027"}
{"id": "train_46566", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "tests_are_truth", "evaluation_metrics", "ci_integration"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "187c40531500ce24e220626003d62a774bb1a339"}
{"id": "train_46567", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "reproducibility", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1742c868a30fd5090fcfa22b7ab97cc702cb7244"}
{"id": "train_46568", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "reproducibility", "evaluation_metrics", "documentation"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "23b1cde87f904d69228457f797401a2f41682302"}
{"id": "train_46569", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "evaluation_metrics", "tooling", "reproducibility"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "93f189aac43befb83356615ed1a4a5957fefd67f"}
{"id": "train_46570", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "repo_scale_reasoning", "governance", "tooling"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "74d8fa29df6d7f7afe5bedfde7fb7d5903b15a7b"}
{"id": "train_46571", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "security_gates", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "768ae4c0b367829d5b635d28fd5a56a1721f678c"}
{"id": "train_46572", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "tooling", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "006bc9ae00ef7e10f967d65898f2f47b8f8352c3"}
{"id": "train_46573", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "auditability", "documentation", "evaluation_metrics"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3eda36850b8fa5020b95cc1cd9f12a23536b8639"}
{"id": "train_46574", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "repo_scale_reasoning", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0164646ede2931d81de8e1166bbaa2670910158e"}
{"id": "train_46575", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "governance", "tooling"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b43c5aafc058bcab5e0ad582ac91590dc3fb5979"}
{"id": "train_46576", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "tests_are_truth", "tooling", "reproducibility"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "92a37f59bbaa9cffee440d96c6712ce44a2768e0"}
{"id": "train_46577", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "governance", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d42162a837ea62d0db6a8b2be46cd6433535f099"}
{"id": "train_46578", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "reproducibility", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b1ac6c0b89c36400037259c7168245f7358c4a79"}
{"id": "train_46579", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "tests_are_truth", "reproducibility", "security_gates"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6998a19a2ea7c54fb040966bd9c7b8b55ae0f0e5"}
{"id": "train_46580", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "documentation", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "383e777deec5ba948a0bc7df7700a5551d91925c"}
{"id": "train_46581", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "governance", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8b259f135fcf248c19ca4152f2c51be0430d8ff1"}
{"id": "train_46582", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "documentation", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9ac6735f9fd47b52756b3761c9c722a0e208628c"}
{"id": "train_46583", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "evaluation_metrics", "tooling"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "97e8a630067669175c2a0d047ceb6b9494d8fa27"}
{"id": "train_46584", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "tests_are_truth", "evaluation_metrics", "tooling"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a189c4af678cb3dd3803e313ac87357c8bdc790d"}
{"id": "train_46585", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "security_gates", "tooling", "ci_integration"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cd28ee1aabfef22b254bc37a551e108f23700078"}
{"id": "train_46586", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "repo_scale_reasoning", "evaluation_metrics", "reproducibility"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c0f19a066bb1282c50e7a68da82d6608edd1abc5"}
{"id": "train_46587", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "auditability", "tooling", "governance"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dc6919e687cd799c96a3786bc051e94f930993e0"}
{"id": "train_46588", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "governance", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c6e57866e027fa3718aa46848036157b4eca66df"}
{"id": "train_46589", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "documentation", "repo_scale_reasoning", "auditability"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "77ca5036363cf09cce5d23622466fa4368697922"}
{"id": "train_46590", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "tooling", "ci_integration", "reproducibility"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f5470e944ce54437a10d01c5ba69eb7b34bde6d7"}
{"id": "train_46591", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "33ce75fd7291f58f4ce6dca5bfa53af4be458c42"}
{"id": "train_46592", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "tests_are_truth", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a7fc4e3507842ea946fc8f763cc31aaee0fa5c88"}
{"id": "train_46593", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["governance", "reproducibility", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3d9ebeac1cbaea7e9d85ca4267df6b7674bcd6c8"}
{"id": "train_46594", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["governance", "evaluation_metrics", "ci_integration", "reproducibility"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "be392eba4a20b071198ead54fb3d0bfa6c90df05"}
{"id": "train_46595", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "governance", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "84d5a7911f92a5ec5372e6a9d89c889b7cddc5d6"}
{"id": "train_46596", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "reproducibility", "evaluation_metrics"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d2fcad05413a410076bae7e518b15feb9f86e759"}
{"id": "train_46597", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "documentation", "evaluation_metrics", "auditability"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "80795618c8466388ba804935faa05ef3b298f84a"}
{"id": "train_46598", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "ci_integration", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5daae6d51f4e128102e6b6b9c9e03a610c76472b"}
{"id": "train_46599", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "tooling", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bd6951e2d6da2fd4a1bd2862df075ccc5f768d67"}
{"id": "train_46600", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "reproducibility", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "93748e94f71390610174fadb834e484f0bc99e62"}
{"id": "train_46601", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5ba6afc90e1a2a61fab3b99e4db76da72a6952b8"}
{"id": "train_46602", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "governance", "tooling", "auditability"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9b71620a408a58b896959960b76b63d9f6a0516e"}
{"id": "train_46603", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "reproducibility", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "72fa2b05cea9907365e309af4c3b3acb674b6101"}
{"id": "train_46604", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "repo_scale_reasoning", "tests_are_truth", "security_gates"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c5255804885dad72b9dbc230b510fbfbee2393f7"}
{"id": "train_46605", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "security_gates", "reproducibility", "governance"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "24816d0e44a6a4cdb571ab0d5fd13a56c2495903"}
{"id": "train_46606", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "ci_integration", "evaluation_metrics", "documentation"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b26fa25dc42dc4ae05398d245c4df84f7de2ba8c"}
{"id": "train_46607", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "auditability", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a4e4c852428f688db3c8100c09cf100bfe4aea47"}
{"id": "train_46608", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "repo_scale_reasoning", "governance"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3686409d01a8eb5efb8dfdd0818f089a2c6410f5"}
{"id": "train_46609", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "security_gates", "tooling"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "13a6cb0a7a8ba5e0d6cae9d2675a99d621978212"}
{"id": "train_46610", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "governance", "reproducibility", "security_gates"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c834ee9cb17ae5f13435905c143471e512145b13"}
{"id": "train_46611", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "documentation", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f8ef4c9c7e4bc62f0b2a812850acaa1482398cd5"}
{"id": "train_46612", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "tooling", "tests_are_truth", "governance"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "68af15e343752f5b5409014e911cc0fd765a3ae8"}
{"id": "train_46613", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "reproducibility", "auditability"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b00be2836bfd6c1b88dae9e5b4d8bc0b8983f7a1"}
{"id": "train_46614", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "security_gates", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b3a08f1dd50f3c377942a20314433fcfbabdb19f"}
{"id": "train_46615", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "auditability", "governance", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "68a694b82aab70114fefe3fd2bbf1cebed6ea09a"}
{"id": "train_46616", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "reproducibility", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "42e963f016de6732bb45009e463d598cf66389ce"}
{"id": "train_46617", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "tooling", "tests_are_truth"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "935958c3ec9bef31e7bf49b877f400d2e34497ee"}
{"id": "train_46618", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "evaluation_metrics", "ci_integration", "auditability"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "26e9ac969a073ca6be22e8e916beaa4c4449ecca"}
{"id": "train_46619", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "568c1791686946f41150092050d56953bf79c7e2"}
{"id": "train_46620", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "tests_are_truth", "auditability", "governance"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2570d1224e6c7e993814c98b2f34f3593928bfb4"}
{"id": "train_46621", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "evaluation_metrics", "reproducibility", "auditability"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "75c074f404d0981f39eea5fc923f0d6ea3c53a64"}
{"id": "train_46622", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "reproducibility", "tooling", "auditability"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e92d8d249386f2c8313cf3bef20a6df1316cbe85"}
{"id": "train_46623", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "security_gates", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bd0f56c60826f90c173b361c4c1449aed7ef2367"}
{"id": "train_46624", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "ci_integration", "security_gates", "reproducibility"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4aa2a38ee97d1b1519e44f2004503d1a588f5e73"}
{"id": "train_46625", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "ci_integration", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0f2560a2f23b0e7cd7c6c423ad44048366d070cb"}
{"id": "train_46626", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "ci_integration", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0d76d0e3bf459fe370808979489c31ea566255be"}
{"id": "train_46627", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "documentation", "auditability", "tooling"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "77984ad8f969f28417f4b0a75da3d02689396a19"}
{"id": "train_46628", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "tooling", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3f3d75f90942c7195f3af59b28c8213eb65a52af"}
{"id": "train_46629", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "reproducibility", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9fe866f1f66a66dd42b4035fe7b00dd57700d334"}
{"id": "train_46630", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "documentation", "reproducibility"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "afa9982da51ccad72c0aeee14dfd0e0b921ec562"}
{"id": "train_46631", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "ci_integration", "evaluation_metrics", "documentation"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8cb9b7cd3f4dffa79709c982e31ee35baaea471f"}
{"id": "train_46632", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "security_gates", "documentation"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2d4b663604e7f54b59f477f7baf8e87474aafe83"}
{"id": "train_46633", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "security_gates", "ci_integration", "governance"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ed8861c6ad1f7ed0366762d975f14d9820fcede2"}
{"id": "train_46634", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["governance", "repo_scale_reasoning", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2851264525351f589538d1ab6ad1d843f12b06b4"}
{"id": "train_46635", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "tooling", "auditability", "documentation"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "838b96baf1b99f5daa3b67f4e6ac9f8246a9454a"}
{"id": "train_46636", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "ci_integration", "governance", "security_gates"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a5a9f172eef06efeeee0f16e5016ec4c0ab86da2"}
{"id": "train_46637", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "governance", "documentation"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c0e216080364cab431a47b77a28bc73f577da022"}
{"id": "train_46638", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "documentation", "auditability", "repo_scale_reasoning"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b0af11fa482f772ae419c241e167094171b09c89"}
{"id": "train_46639", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "evaluation_metrics", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3646a8d8f837be0738e428e7a72c4eaac62dedfb"}
{"id": "train_46640", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "reproducibility", "security_gates"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a89d20396c7a72aee44cc7399a1352099a7a55c7"}
{"id": "train_46641", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "ci_integration", "evaluation_metrics", "documentation"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "913da6160c66f5263f7571b53817a0412c37b71f"}
{"id": "train_46642", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "auditability", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d42831b6e9bfd09597a62b1cbca030bd6cdb2614"}
{"id": "train_46643", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "cost_latency_tradeoffs", "tests_are_truth", "documentation"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4af6048ee7deb648452f9fa5d2cf89bbd44f03c0"}
{"id": "train_46644", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "tooling", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4ce8640026028573b02add4fd0c59a70877c356e"}
{"id": "train_46645", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["governance", "ci_integration", "auditability", "tooling"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5caa926cc2ad543ac1693d69e59cda05a28dd83d"}
{"id": "train_46646", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "auditability", "governance", "ci_integration"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3be63cd75795b58027bcda8c49c6f36779b9cbda"}
{"id": "train_46647", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "reproducibility", "tests_are_truth", "security_gates"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f24610c95e79c6df4024f49446e1ed1c188d97ea"}
{"id": "train_46648", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "tooling", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ab18f5f2d2f09b11147c6ee49fc96dd6a9640e9b"}
{"id": "train_46649", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "repo_scale_reasoning", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ca1f8c24f8a676cccb7963dbdc71471aae4246e0"}
{"id": "train_46650", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["governance", "cost_latency_tradeoffs", "ci_integration", "security_gates"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8243c68c6fd3cb6d2b3ca760bd646ae19b103c75"}
{"id": "train_46651", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "auditability", "tooling", "documentation"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a58271d0d5750b249dc03779502341f440e81e9f"}
{"id": "train_46652", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "reproducibility", "repo_scale_reasoning", "auditability"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eadf96ae058750d41ab65481aa15cacc0a86cc4a"}
{"id": "train_46653", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "cost_latency_tradeoffs", "security_gates", "auditability"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "18c3a39c067612dd1771629f70767e16e3949f47"}
{"id": "train_46654", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "tests_are_truth", "ci_integration", "documentation"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f55c1219d91baf8b1282fb27252cfc38d552783b"}
{"id": "train_46655", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "reproducibility", "tooling", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f1490b53e5c33d604aac137107724b96fc6ac72d"}
{"id": "train_46656", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "reproducibility", "ci_integration", "security_gates"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d97d1e3013199658f81e454395d3c700ab8b0d4c"}
{"id": "train_46657", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "ci_integration", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "27ee7f625efb4a1c58c85f5d44b7cf43f99a5be4"}
{"id": "train_46658", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "tests_are_truth", "auditability", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f8f1b146782a9675477e850958f6d6051da0063d"}
{"id": "train_46659", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "tooling", "repo_scale_reasoning", "documentation"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8db85154d0be67ff7a123d02aa0ed16a4dedb7e0"}
{"id": "train_46660", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "ci_integration", "security_gates", "reproducibility"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c087b4a067b451f249761440e9ebd7185c7354d7"}
{"id": "train_46661", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "governance", "evaluation_metrics", "security_gates"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ba806cfe852c27ebc2839f38d3c335546ed72c2a"}
{"id": "train_46662", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "559ccda68ba4586d79396f350cfd81a350f7200a"}
{"id": "train_46663", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["governance", "repo_scale_reasoning", "security_gates", "reproducibility"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e281a8308799d381215fa2a733e7c1810dd9b67d"}
{"id": "train_46664", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "documentation", "auditability", "reproducibility"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "12d584b6496071761e756e119a6186e350a5b0a6"}
{"id": "train_46665", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "security_gates", "evaluation_metrics", "ci_integration"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e01842931a9b4c42de9acf3746efc6bcb6a9e6c4"}
{"id": "train_46666", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "tooling", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f99d400040ba2aefebb0d72203305cb064ae16d8"}
{"id": "train_46667", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "ci_integration", "auditability"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ca5e5447cbbdf790849d38141c4ff59f07e5756c"}
{"id": "train_46668", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "56a4625d5ce33da2b43abf776e77c604d735d2ad"}
{"id": "train_46669", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "ci_integration", "reproducibility", "tests_are_truth"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "addaa5319fc66b5d7124592c8996e89d04191557"}
{"id": "train_46670", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "documentation", "reproducibility", "tooling"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "26b2466e0fcb42028a5c335594a8efc870b62498"}
{"id": "train_46671", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "auditability", "repo_scale_reasoning", "tooling"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fc28c0bd073533976a336f78be52cfee7c07c2d0"}
{"id": "train_46672", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "tooling", "documentation", "tests_are_truth"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "766ad2309c65e447a8ed69f087fe3e60ca47e6ec"}
{"id": "train_46673", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "tooling", "documentation", "auditability"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e290b0e8376806a2aff1ee98e12913cf4eed3b00"}
{"id": "train_46674", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "cost_latency_tradeoffs", "governance", "ci_integration"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "39addb71358d353d0237e7c5cf4e49fe97ee6e54"}
{"id": "train_46675", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "governance", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ab04b7223c53a6a788b7f6735d3798dce2e313a7"}
{"id": "train_46676", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "documentation", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "39237b43e9c469f525bd284cef329033e3fd7386"}
{"id": "train_46677", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "governance", "documentation", "tooling"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1f4355d5173e126a317eeb39f4347d1e57c9c0a5"}
{"id": "train_46678", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "documentation", "ci_integration", "tests_are_truth"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aaa959ad798b512bbcc47e8f22c4db27ff206b79"}
{"id": "train_46679", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "repo_scale_reasoning", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "805b8154faddbf1963a15a64433bf336d86e4ec7"}
{"id": "train_46680", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "evaluation_metrics", "security_gates", "reproducibility"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8aea59f6a0e7699dafe1dd8ce240f07cc89bf971"}
{"id": "train_46681", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "tooling", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8691f84e940d513acdbc69087dcd2e55d1deb5a3"}
{"id": "train_46682", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "tooling", "governance"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "db643c123df2f94028cf593a78c4f79116d0f946"}
{"id": "train_46683", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "tooling", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f171040221edc5cc584ef9a066c3ccb4cb5395da"}
{"id": "train_46684", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "reproducibility", "security_gates", "auditability"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc7d2ca2f5b1b2e456eed1fff7521a430d6f2595"}
{"id": "train_46685", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "governance", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5f8096097e3adc767d0c00534b4d12578e8230f1"}
{"id": "train_46686", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "documentation", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0c515e0b14d151b1070ed8cdd2aa35d82b2429ae"}
{"id": "train_46687", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "documentation", "security_gates", "tooling"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e2acc537e3921c4d9e867eedb193e515b50fdb33"}
{"id": "train_46688", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "governance", "auditability", "tooling"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5397acbcc7b86ebce03c5343d6d1d24b2e183552"}
{"id": "train_46689", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "governance", "auditability"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "85c3892a77251ab1a47d6dd7da37253eca1ab715"}
{"id": "train_46690", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "documentation", "repo_scale_reasoning", "security_gates"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4616537a9fa0421e4d0cdea3bcbbfc25d845c18d"}
{"id": "train_46691", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "security_gates", "auditability"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "daeef27260d2c2696edf1b1bfaed113bbc11b179"}
{"id": "train_46692", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "reproducibility", "documentation"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c9d676d22ddb713671adbc32d299dc6ff319205a"}
{"id": "train_46693", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "reproducibility", "evaluation_metrics", "tooling"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f621e30bd75e0958c61a42d5a4b8627f8c1219be"}
{"id": "train_46694", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "security_gates", "governance", "auditability"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5a5c21dce1a2ed0d97bce98c3ea416adea9d3b78"}
{"id": "train_46695", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "evaluation_metrics", "documentation", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d1e4459347b20020552aaa7b57d836300b6e6c4c"}
{"id": "train_46696", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "tests_are_truth", "ci_integration", "reproducibility"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0f6c7656453f0a74e86f1e41b2711df0255ee140"}
{"id": "train_46697", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "ci_integration", "repo_scale_reasoning", "governance"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1a4263c786230699a46fa9295874d71104c487b5"}
{"id": "train_46698", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "governance", "documentation", "reproducibility"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3c67a039c5dc626f7703afdb5dbbd1917e9379a6"}
{"id": "train_46699", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "tests_are_truth", "reproducibility", "documentation"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fbecaa05200dcd02ab78b6422677f7b29f37f3f4"}
{"id": "train_46700", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "documentation", "tooling", "reproducibility"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "14aae19b17f3bfcf4abb7c6002f2888f3d1bc3ac"}
{"id": "train_46701", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "governance", "auditability"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "14992629785714094062c32a5a15c0b96c8b2d2a"}
{"id": "train_46702", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "reproducibility", "governance", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2084a496545a43f20c292864964e74e69f49af0f"}
{"id": "train_46703", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "ci_integration", "reproducibility", "evaluation_metrics"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "87dfab0b64d657b78a59d915c03f6d26a8fa7cef"}
{"id": "train_46704", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "cost_latency_tradeoffs", "reproducibility", "documentation"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7095dcdbd18c49bdbf1983dde81790fc12f87151"}
{"id": "train_46705", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "security_gates", "tests_are_truth", "tooling"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "712f1790f1e386981c88d1c32d94dde20b51fb8b"}
{"id": "train_46706", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "governance", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "efffe8da8e3be507c5eb872fa633a47179843c01"}
{"id": "train_46707", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "reproducibility", "auditability", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c48214e604a8c291d682fba4320d17f59b0ddc40"}
{"id": "train_46708", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "security_gates", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fa7c839028ce73706e855694703e5b999c5755a8"}
{"id": "train_46709", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "tooling", "auditability", "documentation"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dc5cc4f03e142e08481d0f53dd2b4ed9d630e71d"}
{"id": "train_46710", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "security_gates", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eb10dd3d892dcfae7c2505d6f8f2e7011cae3016"}
{"id": "train_46711", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "tooling", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "99d6cc95d74008f61c697773d691ee87e75732f4"}
{"id": "train_46712", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "repo_scale_reasoning", "auditability", "evaluation_metrics"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a2bb68bb34b6749ffdd4922fa2d6c577d4f13d27"}
{"id": "train_46713", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "cost_latency_tradeoffs", "evaluation_metrics", "tooling"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ebdf2da374b96f062706d6b58f048b0468ec5a95"}
{"id": "train_46714", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "tooling", "documentation", "tests_are_truth"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "28b249ba44949a1cc04c8f87a693512b11a5542d"}
{"id": "train_46715", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "security_gates", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c4617ab8451239592564ecdd6d79a0f936a71df6"}
{"id": "train_46716", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "security_gates", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b3295b3c6c1a24c3f5618f0422b6f57cb8601790"}
{"id": "train_46717", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "tooling", "governance", "evaluation_metrics"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1c75d41f9117a12708f3b0f1f273cf54bd86a7ee"}
{"id": "train_46718", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "security_gates", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "42294db117eb5ba876939dd3e8994b3275ee80d9"}
{"id": "train_46719", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "reproducibility", "ci_integration", "auditability"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c7ed5cf9c0c084f399d7eafc9749d5de38ce321c"}
{"id": "train_46720", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "reproducibility", "tests_are_truth", "governance"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c8be278f72e10a7ca7c85c0f8ef27b901269f8c1"}
{"id": "train_46721", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "cost_latency_tradeoffs", "tests_are_truth", "tooling"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1ddb6d41bb5358c3602256d43f646a4732a08beb"}
{"id": "train_46722", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "reproducibility", "auditability", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "006bb37b57654e404c47841a60080bee06d3c9f3"}
{"id": "train_46723", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "evaluation_metrics", "security_gates", "governance"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d0a27b9298748f87f9b797a9dfb4a2491b5cd80f"}
{"id": "train_46724", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "reproducibility", "ci_integration", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e620f1c64a2c8a61ec049e793447e567ac313bb7"}
{"id": "train_46725", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "documentation", "tests_are_truth", "ci_integration"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "21519f406f13525fc57f5f656a5f5cb449381152"}
{"id": "train_46726", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "tooling", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0ffea5d4416f3e830e91a1726a38147f7909b549"}
{"id": "train_46727", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "repo_scale_reasoning", "tooling"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "15d431c6b8d5685b7502511d54d0689ac6a7d7bf"}
{"id": "train_46728", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "documentation", "governance", "evaluation_metrics"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ba0070915dbc38fcab39d86f202689e203e4fe7b"}
{"id": "train_46729", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "documentation", "reproducibility", "auditability"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "31718150a2a6c5e55844e150dd959f06bd0a697f"}
{"id": "train_46730", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "ci_integration", "documentation", "reproducibility"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b20ed7e02913654823ddfce90bf117390e8c6af6"}
{"id": "train_46731", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "reproducibility", "security_gates"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "454bd856ec967172b65c62514b408457b09a8cac"}
{"id": "train_46732", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "security_gates", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "995f0dab2c7b6e09ed7994fbd1dcfe772e3d931d"}
{"id": "train_46733", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "repo_scale_reasoning", "security_gates", "reproducibility"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3add3d6fb41910e951d117dbd74ef92a33a1041f"}
{"id": "train_46734", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "auditability", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "357289dedfa6831c79c6292d23e8bae32d210085"}
{"id": "train_46735", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "cost_latency_tradeoffs", "security_gates", "governance"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7471989bc81435145ac9620ec2850bebea228ec0"}
{"id": "train_46736", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "repo_scale_reasoning", "governance", "documentation"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7d00a3d5080150cfd3d2b20a47ad4266bc0e6d26"}
{"id": "train_46737", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "tests_are_truth", "ci_integration", "security_gates"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e487d0618e89ec583486a6f48c8f59734f5defef"}
{"id": "train_46738", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "reproducibility", "ci_integration", "documentation"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9c4d12cab46fd7aa4a6b16e68c8c16e5c2935b5e"}
{"id": "train_46739", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "security_gates", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "25fc3166ef5e14907d2e7b1caa8bc500c8c0e2f9"}
{"id": "train_46740", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "evaluation_metrics", "documentation", "tests_are_truth"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "97d10b0830c485492275a72917d6edac8c31ab18"}
{"id": "train_46741", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "reproducibility", "auditability"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "500053cea39fa7057c14f56c10cfbcdee84e6d98"}
{"id": "train_46742", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "documentation", "reproducibility", "governance"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "52a634069a9799c4d7d2b247f6385543fa6c5e68"}
{"id": "train_46743", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "tooling", "tests_are_truth", "auditability"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "840d7eb46c90aa7c7e973b98cf6e824771861fed"}
{"id": "train_46744", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "evaluation_metrics", "cost_latency_tradeoffs", "governance"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fba2a8da22d45a45214ce5320afb12c81778ca45"}
{"id": "train_46745", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "evaluation_metrics", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1c9066634eee990e8ac31473807d0d621e02f57f"}
{"id": "train_46746", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "tests_are_truth", "auditability", "documentation"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "01c22da15e74ba1a072aeb079125346503c3c1a4"}
{"id": "train_46747", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "documentation", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3105245b56218c5b4ab47a6774e768972513dac2"}
{"id": "train_46748", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "security_gates", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b49827186dd657fa281ca0d5aeb76251c2196e34"}
{"id": "train_46749", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "tooling", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f84717487db315236cc9e690a6e3601d0d672200"}
{"id": "train_46750", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "governance", "tooling", "tests_are_truth"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c45fbb2b432e148f0f94c3cf93c4bcd587ad86c0"}
{"id": "train_46751", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "evaluation_metrics", "auditability", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "12e5d857cdc9c882fdaa260f15da3fa1e68b8acc"}
{"id": "train_46752", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "tests_are_truth", "reproducibility", "governance"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a9bfcd7675aad532cc4175b5a791fd88cdd83028"}
{"id": "train_46753", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "repo_scale_reasoning", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ef980fffcbb60bfc2310e9b2e18ef8e838b8989f"}
{"id": "train_46754", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "tooling", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0cc582385ff569c0f8e90dca0f97474c48be578f"}
{"id": "train_46755", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "tests_are_truth", "ci_integration", "security_gates"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d8bf04dd956bbae7bf036c184e72681b636162e6"}
{"id": "train_46756", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "governance", "auditability", "tooling"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e424cccc5cf3eb7212823c62a17cf96c81c63071"}
{"id": "train_46757", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "reproducibility", "tooling", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "db2d551cd80957ff0c52f9638e63a6ed3618a381"}
{"id": "train_46758", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "security_gates", "evaluation_metrics", "auditability"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c52c22eb9186c34162533b9e84e810a136d9aa82"}
{"id": "train_46759", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "evaluation_metrics", "security_gates", "tooling"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2e893ba5e349de855d2a1f8ed75edb1162b9f449"}
{"id": "train_46760", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "documentation", "auditability", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4a01dac3255fe0f380672842133a95f7beed8fdd"}
{"id": "train_46761", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "auditability", "documentation", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cd77e13b7ed685e1fe01b2710e230792325b24f0"}
{"id": "train_46762", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "security_gates", "evaluation_metrics", "reproducibility"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c023aa0e8b1ba6ca3937d8b0dbc11ba86b5f00b2"}
{"id": "train_46763", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "tooling", "auditability", "security_gates"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8ff5b3effb0a1a3e2ab3d79cd466f148978aa43f"}
{"id": "train_46764", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "tests_are_truth", "security_gates", "documentation"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b04f48796c167bbbdd4cdff6872c0cf6ad4c56cd"}
{"id": "train_46765", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "governance", "auditability", "documentation"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "277dca203bf1c3e829e9e5411b487d90a968a52c"}
{"id": "train_46766", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "governance", "tooling"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "89d0aa5e354e5195c3a8730aa59e1ae467940051"}
{"id": "train_46767", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "evaluation_metrics", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5e2e837f64dc8163c4ca44d0e3c28ad4ab6ceb95"}
{"id": "train_46768", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "cost_latency_tradeoffs", "tooling", "evaluation_metrics"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "44f5e80d0db7435c849aed0e151e5d07f186d92c"}
{"id": "train_46769", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "documentation", "ci_integration", "security_gates"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a7892df77cb6e6ac5d922b8509952302583eb1ab"}
{"id": "train_46770", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "auditability", "security_gates", "tests_are_truth"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "53201ddb02d1bcadd1c522035cb9f92066226552"}
{"id": "train_46771", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "reproducibility", "tooling", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "44e64ab400e8d784a92637a92a4e2d6417889fd9"}
{"id": "train_46772", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "governance", "repo_scale_reasoning", "auditability"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0596fea7e9bd2bdd7c952d2e10eb78f553df59c0"}
{"id": "train_46773", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "repo_scale_reasoning", "governance", "evaluation_metrics"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0fd8a3a3c11c6cb07ebfe8d83a2ad1ac18214e33"}
{"id": "train_46774", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "tests_are_truth", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6daa5b26669ff306ce763f71aecaaea06b1365f7"}
{"id": "train_46775", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "governance", "security_gates", "documentation"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2f30244dd7159d32d78242629de68a434cd4aad0"}
{"id": "train_46776", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "security_gates", "tests_are_truth"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b0aa8489ad9882b7711f2bc150f03fcedff28ebb"}
{"id": "train_46777", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "governance", "reproducibility"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cf925f20c09ead6074683f817c212e3abfa5aa20"}
{"id": "train_46778", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "security_gates", "governance", "tests_are_truth"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d39994ac566175c227631e77ee9252a9c4611224"}
{"id": "train_46779", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "ci_integration", "auditability", "documentation"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a163d7556d6475c45e5ebcb5eca72df834c0238e"}
{"id": "train_46780", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "security_gates", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cc7e524446e7a9f0a8fc28a20608040cc3733c63"}
{"id": "train_46781", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "reproducibility", "documentation", "governance"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1de4eb99959a5ee92c6ee32c2388dfe388f62a39"}
{"id": "train_46782", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "documentation", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ca62baf89cdef451610b94a544971678635c14f3"}
{"id": "train_46783", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "reproducibility", "evaluation_metrics", "security_gates"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a8c3ee8e052abe1635f0ae59db99c6073ff08abf"}
{"id": "train_46784", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "auditability", "tests_are_truth", "documentation"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dd298fd15ba61c7dc8b967a256c9f4c31305c638"}
{"id": "train_46785", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "tests_are_truth", "tooling", "repo_scale_reasoning"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d1c9c0c530434a1e3c13f07e699d09f379fde63b"}
{"id": "train_46786", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "evaluation_metrics", "documentation"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "498a9d816c69025ffcf443fe9e044b44a0d31c0e"}
{"id": "train_46787", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "governance", "repo_scale_reasoning"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a12f13f1f11e92859e6af45aa5c14c5fed079260"}
{"id": "train_46788", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d7cd44bc799f19f35b5b64b418fb4b1857c4adbd"}
{"id": "train_46789", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "ci_integration", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "31119fc39547d5d1ffe9b50fecd13dad057b9352"}
{"id": "train_46790", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "governance", "auditability", "tooling"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7dfb7a257d9f5f11b51c38ef89c86e4be0ecfbb9"}
{"id": "train_46791", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "governance", "repo_scale_reasoning", "documentation"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cc72c192bae800a8215212cc82b57ea7c09ea2a2"}
{"id": "train_46792", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["governance", "documentation", "auditability", "tests_are_truth"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d0ac185bc57e01b8c321b5fffe3845bb9ca8857e"}
{"id": "train_46793", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "cost_latency_tradeoffs", "auditability", "evaluation_metrics"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "600fd9e7554332a79b655f71c51f232bdb39d8ec"}
{"id": "train_46794", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "auditability", "governance", "security_gates"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "44edcca5bdfe6dc0e2f4f9692f89ca58e11764f9"}
{"id": "train_46795", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "reproducibility", "governance", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "091641a2f464359ce15e22aad700510ace19feed"}
{"id": "train_46796", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "repo_scale_reasoning", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "82450652cced7a93d439e9853e54a2902d329e53"}
{"id": "train_46797", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "evaluation_metrics", "tests_are_truth", "tooling"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bf94109358a5c5f7c731210c9c6b7ef39c2aa4b6"}
{"id": "train_46798", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "governance", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e6544cad276c07e214ed4d1eceda9b4597e294b8"}
{"id": "train_46799", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "ci_integration", "security_gates", "documentation"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "016a8df867549d7005d299c033ea5409b45c8de6"}
{"id": "train_46800", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "governance", "reproducibility", "ci_integration"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d9d7f7ebf7d766e3ead93b68f61e9169ae753701"}
{"id": "train_46801", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "tooling", "documentation", "auditability"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9acac950b622f8ade24a8b56fe52ef55c3a153c6"}
{"id": "train_46802", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "repo_scale_reasoning", "security_gates"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "17b759c8f43062cb20b086efee53f63da9867f83"}
{"id": "train_46803", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "repo_scale_reasoning", "documentation", "evaluation_metrics"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ca04357f9ad03ca6d8b7657b7a35c4efd49e6f17"}
{"id": "train_46804", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "documentation", "governance", "tooling"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b49827186dd657fa281ca0d5aeb76251c2196e34"}
{"id": "train_46805", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "ci_integration", "documentation", "security_gates"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "11d3e256948e6a888d14f8fc823a84ac9b073271"}
{"id": "train_46806", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "cost_latency_tradeoffs", "security_gates", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "40b49d2e62a1d649fa71b7e5d297cec5259365e4"}
{"id": "train_46807", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "tooling", "auditability", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cfef7004ec4d0f43acf36b0c9d8fb6353795aa39"}
{"id": "train_46808", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "reproducibility", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9394d90b29e69f78ea45c913297c6a0c96867efa"}
{"id": "train_46809", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "cost_latency_tradeoffs", "governance", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a4e436ec2c70ae32be0b26c7d1c675857e26aa34"}
{"id": "train_46810", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "documentation", "ci_integration", "tooling"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3b239e94dc1c2bc462325b01b6a26166cc827c69"}
{"id": "train_46811", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "ci_integration", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "286f3e2bf7c4546eddd8ac4289fa1113697528da"}
{"id": "train_46812", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "documentation", "tooling", "tests_are_truth"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1c6f651767365db0bda4887959c5bea19d047f13"}
{"id": "train_46813", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "evaluation_metrics", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "379eb6b2ba9932c9c10a6da5097bdbb239dbb9d7"}
{"id": "train_46814", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "ci_integration", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4a46e7aa438f06bc2ba9458355a88ab061f9327f"}
{"id": "train_46815", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "ci_integration", "security_gates"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7115c87e1e14682e0e47af6a7bf4125432a2b2bf"}
{"id": "train_46816", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "ci_integration", "tooling", "reproducibility"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3709b4cffec0c76f1cac074cf25d2f2683a023f2"}
{"id": "train_46817", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "auditability", "security_gates", "ci_integration"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "36ccf8d160aa32177598b80e43ede6e05e6543cc"}
{"id": "train_46818", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "documentation", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7513d962fe005ba81a332ff7a1690233d043bc8a"}
{"id": "train_46819", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "auditability", "tests_are_truth", "ci_integration"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2623bdda273fee4c1d9b0e3f951f2ffc572364dc"}
{"id": "train_46820", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "cost_latency_tradeoffs", "tests_are_truth", "tooling"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f413e73d7821a0d3e2557a3e4d8b029e21a7224e"}
{"id": "train_46821", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "governance", "evaluation_metrics", "ci_integration"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3af7ecf7c569fe65c824205cfa14158744255790"}
{"id": "train_46822", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "auditability", "governance", "tooling"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "90ef92d0f50336c1d1624b0fb09dddf73cef6452"}
{"id": "train_46823", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "auditability", "governance", "tests_are_truth"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7b45efbc172abf1c457e2a92f69233c508941442"}
{"id": "train_46824", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "tests_are_truth", "reproducibility", "ci_integration"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5ee0d4e6e41056bf6101881a21759fba51ad156d"}
{"id": "train_46825", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "governance", "auditability"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fe07213fdbf5a531c25bf2c6f2d732da21e391f4"}
{"id": "train_46826", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "documentation", "governance", "reproducibility"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "be15e2cc4b72ffc14ab16d872a2fdd1475796067"}
{"id": "train_46827", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "evaluation_metrics", "ci_integration", "tests_are_truth"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3689fa4c5740e3fcf38e67774bcbd36fd6efad8c"}
{"id": "train_46828", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "governance", "tests_are_truth", "tooling"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "41a04286dc87d29842a6ee17cde4d3829b98af5b"}
{"id": "train_46829", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "reproducibility", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "22b174554b97ff17d1de2d9848f59a6d5a14572a"}
{"id": "train_46830", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "documentation", "governance", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ba07f3b02ff71b3cf4abe193794831124f782b68"}
{"id": "train_46831", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "repo_scale_reasoning", "tooling", "security_gates"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f8e6b02933e390c8c06807d1061d478fa0435fca"}
{"id": "train_46832", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "security_gates", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "40e767972bd63a1ea2bf97ad7ed58f5da366f390"}
{"id": "train_46833", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "security_gates", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a7746af5b39c606c7bf36bfb082800d0d4085911"}
{"id": "train_46834", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "reproducibility", "evaluation_metrics", "ci_integration"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a84f781f10b60ecf95d4e0b12671eae547bb8ab1"}
{"id": "train_46835", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "cost_latency_tradeoffs", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "99d6acf94a77c30ef96d6c6caac297908f7f6208"}
{"id": "train_46836", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "governance", "ci_integration", "reproducibility"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0eabf520cea127a29430e5f4e385f2c8d6b4f193"}
{"id": "train_46837", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "repo_scale_reasoning", "tests_are_truth", "ci_integration"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3e1e267e21101f0ccfa5148e5d89749d1e8f3794"}
{"id": "train_46838", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "repo_scale_reasoning", "security_gates", "documentation"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ccd58af80d2f79ac38ad423600d1c784e7b92a99"}
{"id": "train_46839", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "evaluation_metrics", "tooling", "ci_integration"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "82e5b08df66b86992d0749d22a660a41cf5048f4"}
{"id": "train_46840", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "ci_integration", "tests_are_truth", "security_gates"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "610541442ffd2b1f1c3f2e008d3d1749ab5af6c7"}
{"id": "train_46841", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "ci_integration", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "96248cf5e9d92f7d41d6121d743a5b1f19e0cb3c"}
{"id": "train_46842", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "reproducibility", "tooling", "evaluation_metrics"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "011ce5de6e95675c5f64d64fc6f317c1ecb7921e"}
{"id": "train_46843", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "evaluation_metrics", "tooling", "reproducibility"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "730e7ab9ef0204eb8dcb667591ae85b8b3e7150c"}
{"id": "train_46844", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "reproducibility", "governance", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1bdb886286cf3a923b23afd0dbe7bfa2f8368941"}
{"id": "train_46845", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "ci_integration", "tests_are_truth", "tooling"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "07734809fb74b9a7a8741fd4a52c5238e7fb6589"}
{"id": "train_46846", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "reproducibility", "auditability", "tests_are_truth"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cb46b44a5bdee05a218ce561055af994be992d53"}
{"id": "train_46847", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "governance", "documentation"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3685f6677b01bf21e5ea5c691b6ca38c72379d24"}
{"id": "train_46848", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "evaluation_metrics", "ci_integration", "auditability"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b74c8a91f3861e9204c93c908640dca4e419d932"}
{"id": "train_46849", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "documentation", "reproducibility"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a7188ee3cb0de5721e815a26041eb4992a7e14b4"}
{"id": "train_46850", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "repo_scale_reasoning", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "407a87549d82ad5ee08e2a55357daad561f7ebcf"}
{"id": "train_46851", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "tooling", "security_gates", "documentation"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "de37c771ebf8f6fcad5b461d88324de45e48eb7f"}
{"id": "train_46852", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "auditability", "ci_integration", "tooling"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "43fcfd556d5e00182c71b2736a81f26e875f0681"}
{"id": "train_46853", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "tests_are_truth", "tooling", "auditability"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c5cb7c1c54d56d710154c2ab25afd23a45049ef0"}
{"id": "train_46854", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "tests_are_truth", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3b3d396a1f79e41bf3620bf515f19db51dcb7d9f"}
{"id": "train_46855", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "auditability", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "413d45c93028172886eeee07bd963e560aa000fc"}
{"id": "train_46856", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "security_gates", "tests_are_truth"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "97bd8831c291610e35274ce4305e061407ad0335"}
{"id": "train_46857", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "auditability", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "655d9fcd55ac7a254feb7c7f32367d05b6151dda"}
{"id": "train_46858", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "security_gates", "ci_integration", "reproducibility"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "425a88cb10f2d4818159eb7349d208dc3d9340bb"}
{"id": "train_46859", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "governance", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b376fe5f453450e80fd3eb7980cb573edf008332"}
{"id": "train_46860", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "auditability", "tests_are_truth", "tooling"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "75adbc8e78274a5f19ff95fad2d764dd7834cc1d"}
{"id": "train_46861", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "governance", "ci_integration", "security_gates"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "544bc27f1f5222ffb8509ecb7a2a6b6f4c906afa"}
{"id": "train_46862", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "ci_integration", "auditability", "reproducibility"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1f7d5487ea2d108b8f0ee32e1f59447a30e1b213"}
{"id": "train_46863", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "reproducibility", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b60acde04ac653389b01402d11cc5ea8f6531c7b"}
{"id": "train_46864", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "cost_latency_tradeoffs", "tooling", "reproducibility"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cd39546701d6ae6f004373810ee99102b9c67be1"}
{"id": "train_46865", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "auditability", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d518a25db85ff5bfa817c474c9c5b93fd3964858"}
{"id": "train_46866", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "tooling", "reproducibility", "documentation"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bdb7fe1e39859a31182875d0419887c111db9de2"}
{"id": "train_46867", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "auditability", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7ee5a05912df57f1fb8cb738d6bee5c04ef09c24"}
{"id": "train_46868", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "documentation", "reproducibility"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "43a19ca5bdba388b1961b4796070648e90f10b7a"}
{"id": "train_46869", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "evaluation_metrics", "tests_are_truth", "tooling"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "46c8fe37d3fd320156c5cdc823dddf1a607cee1b"}
{"id": "train_46870", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "security_gates", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0fc489605312c6ef83c18c61bfb053e7abbd9189"}
{"id": "train_46871", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "ci_integration", "tooling", "security_gates"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5993ad71f417e23d6dd64453be4236defd0de693"}
{"id": "train_46872", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "reproducibility", "evaluation_metrics", "security_gates"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ea8a235f197e0ab16d75ca79821a794019b6423c"}
{"id": "train_46873", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "tests_are_truth", "governance", "security_gates"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f11136efff252704374d1467638a3530f52fffb1"}
{"id": "train_46874", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "auditability", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c7c15bbda525a7fcbadb92270d7735aaceb45e09"}
{"id": "train_46875", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "tests_are_truth", "reproducibility", "tooling"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ee24ba7d50dd584b68027a783a8962921140ca72"}
{"id": "train_46876", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "tooling", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f1c31949bd9519f5d8c46f7ed63a1977fbe3e1eb"}
{"id": "train_46877", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "security_gates", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fdc06aa1eaf1bec5750c2501758994e86aea3504"}
{"id": "train_46878", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "repo_scale_reasoning", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6e8994b925c191aa94fc9c158a576d902d9c6e73"}
{"id": "train_46879", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "ci_integration", "reproducibility", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5ba2d0a2002511a47c34a3d2f8aeaf0691a00b0a"}
{"id": "train_46880", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "documentation", "governance", "security_gates"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "77be0753357ff2dbff9081fb7fcce13f9363f7a3"}
{"id": "train_46881", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "security_gates", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "89c072a67e24b89dd9237a6d1669891105299965"}
{"id": "train_46882", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "auditability", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9826152a14acb32acce17b80cd468dba226f0c87"}
{"id": "train_46883", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "evaluation_metrics", "ci_integration"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "660456686fc1ef7c5808098d3758cf26ab43fc05"}
{"id": "train_46884", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "cost_latency_tradeoffs", "repo_scale_reasoning", "governance"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5712a21ec9f76150dc6c79b673984f4f794b2f80"}
{"id": "train_46885", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "ci_integration", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9b97134354e80d1227e7161fff7f09b429a8c71c"}
{"id": "train_46886", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "documentation", "tooling", "ci_integration"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4a1ce4ef0088125a37ffaec943bdb92d6a0b2eb8"}
{"id": "train_46887", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "security_gates", "documentation", "ci_integration"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6ed63d2e92b725c261cdf56338923254c71cde46"}
{"id": "train_46888", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "ci_integration", "tests_are_truth", "tooling"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "44071079d322bc52d2ef47083940e25ed6ebe499"}
{"id": "train_46889", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "tests_are_truth", "reproducibility", "documentation"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8b299289ecc4cad3a8af417aa7df23e9ee308b8e"}
{"id": "train_46890", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "governance", "security_gates", "auditability"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "176c629c493c086125d10377e7889801f2e22161"}
{"id": "train_46891", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "repo_scale_reasoning", "ci_integration", "tooling"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1ced8645d458f12c1cbdd27f87896fc9ff1da887"}
{"id": "train_46892", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "repo_scale_reasoning", "security_gates", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e4da684296a85562912b6dd5f291eac39234d20e"}
{"id": "train_46893", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "reproducibility", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "81667b547f60eaf54634a4671061f651e407be5d"}
{"id": "train_46894", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "cost_latency_tradeoffs", "tooling", "evaluation_metrics"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "825ed42826a2d98700b74b1da6df182d6251d8ed"}
{"id": "train_46895", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "security_gates", "governance", "auditability"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e678393b446d6979b1dce6f3a05f7051f5db3cb6"}
{"id": "train_46896", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "reproducibility", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5b64a8502e5738ced9375d1c78c44de897b4c546"}
{"id": "train_46897", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "reproducibility", "auditability", "ci_integration"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eff70011e6b0c03b0e6403bbfad273b9350ffa5a"}
{"id": "train_46898", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "auditability", "security_gates", "governance"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "58081009a991edf8430e9588c87a03eae54ec29f"}
{"id": "train_46899", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "reproducibility", "governance", "auditability"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8e59a788608e37f6019199c497704f75572de0e7"}
{"id": "train_46900", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "reproducibility", "tooling", "evaluation_metrics"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "155c3cbda080b7c715b6854bd17281da9e19dc59"}
{"id": "train_46901", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "documentation", "tooling", "ci_integration"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "53c78dd427b90de500938e9d5d064889ccbfafce"}
{"id": "train_46902", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "tests_are_truth", "security_gates", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dafda0f5e7844ac27f64e732cf97308ea99e51d5"}
{"id": "train_46903", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "evaluation_metrics", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dbf5c95d57d1bff4a3d6ecc96cb44571a8183868"}
{"id": "train_46904", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "governance", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a0b4f921a6ea118f115b2e795e874a3a47921d0c"}
{"id": "train_46905", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "reproducibility", "governance", "documentation"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b353c2913e5d701928871a8bd4cb8a1a6c4ef16f"}
{"id": "train_46906", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "tests_are_truth", "ci_integration", "reproducibility"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f7dbd87e4e7b8fcd871adc256d833f3aa4bf5d08"}
{"id": "train_46907", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "documentation", "reproducibility", "governance"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b9a5fa06f23e65e106deeace7f7937750324d555"}
{"id": "train_46908", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "auditability", "tests_are_truth"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e0a85391fef5c8e3f8c2cbf6921151e6ad966a6e"}
{"id": "train_46909", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "documentation", "governance", "tests_are_truth"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cc79e15751ec4c5d36964e4aa9955d452b1ee631"}
{"id": "train_46910", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "tests_are_truth", "governance"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2de728827fc60559fcc51d53138b928c467346ec"}
{"id": "train_46911", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "reproducibility", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "42ca21e67b76fb6d442af95ce756ad1f33bed0c8"}
{"id": "train_46912", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "tooling", "auditability", "reproducibility"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e609480a1c7f2ac2fb74d6deef21a664a5b05610"}
{"id": "train_46913", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "security_gates", "auditability", "tooling"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "095bbf83e46660b80fa7806c41ed2e5a8ead70a9"}
{"id": "train_46914", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "evaluation_metrics", "governance", "auditability"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "56847a90fe8b0ff8b828121106d97092690082d7"}
{"id": "train_46915", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "tooling", "auditability", "tests_are_truth"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f6522bb527e8bb97779def2debe1a7cf5c47ca78"}
{"id": "train_46916", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "auditability", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ab1c704018fc1aa4fdb4b461f3665ea931cd9437"}
{"id": "train_46917", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "tooling", "ci_integration", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c0797a96fbe2f5093d7f94f3e32d953ae41b5198"}
{"id": "train_46918", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "tests_are_truth", "ci_integration", "reproducibility"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c9dad03f193210a4f2dc82f2d22a7944d857383e"}
{"id": "train_46919", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "auditability", "reproducibility", "tests_are_truth"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b439304067b509063461b79ad10e921c97a23074"}
{"id": "train_46920", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "ci_integration", "security_gates"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a2de4c8684a9443f5b0d7e3585aa49b60f923c80"}
{"id": "train_46921", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "tooling", "governance", "auditability"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1ff161c8559e2292f8306c7fd25ad1bc0850eff7"}
{"id": "train_46922", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "evaluation_metrics", "repo_scale_reasoning", "documentation"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0aea0f1f70f0c8000e4095d0fc3336c1864b2bda"}
{"id": "train_46923", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "documentation", "ci_integration", "auditability"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b34220b2217e55ab0eb53c7a11f9ac642ae521ef"}
{"id": "train_46924", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "evaluation_metrics", "documentation"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b7d35199d022253629a02a87a3ad5ba75e7548bd"}
{"id": "train_46925", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f42b921c5c9d6eed4dc0a89ff3482228ebf5a0e4"}
{"id": "train_46926", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "documentation", "tooling", "security_gates"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "253a05b716a12460724a5f0d69076cfc7c62989d"}
{"id": "train_46927", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "documentation", "evaluation_metrics"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4617ac7f1948260168dbb7fcf38a5e971e163220"}
{"id": "train_46928", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "tests_are_truth", "repo_scale_reasoning", "documentation"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d3a3f3b2ff97fa6f8d590d6bf8eb8ee0a6e17841"}
{"id": "train_46929", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "cost_latency_tradeoffs", "ci_integration", "security_gates"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fd344743fca8d1b85ed986602802f2ae0c63cdef"}
{"id": "train_46930", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "tooling", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3ab9287879cbb1bec898d6ed2c8cc2d532f3647f"}
{"id": "train_46931", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "tooling", "reproducibility", "tests_are_truth"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "681f926274a12f5e7eb9ae9566c8ce22c7b86adc"}
{"id": "train_46932", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "security_gates", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6011afc3016aeb721a0d3d3a9cc4921ebbb45a03"}
{"id": "train_46933", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "tooling", "repo_scale_reasoning", "security_gates"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e5307a9bf15eda34502890239eed11f8181731ea"}
{"id": "train_46934", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "security_gates", "tests_are_truth"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "95de2a4fec0d93290a2889cf575a98820e432e5b"}
{"id": "train_46935", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "reproducibility", "security_gates", "auditability"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cac780c28178d9fada05067fc20f6b4259f4849f"}
{"id": "train_46936", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "tests_are_truth", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a22a9a7eb4bcdfe8f5de5b8f9e435cf151975ad0"}
{"id": "train_46937", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "evaluation_metrics", "ci_integration", "reproducibility"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e7248a348e7c624e4dae08d2ac1bbde8f7379c27"}
{"id": "train_46938", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "ci_integration", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3cca022a649cfacd5c5da169269a70faf7ca1cad"}
{"id": "train_46939", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "repo_scale_reasoning", "reproducibility", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "49215e6c7fb5a554bd1c072c802424182587d8d8"}
{"id": "train_46940", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "tests_are_truth", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c4298976b520c9481abaadc11882fc4b944e7701"}
{"id": "train_46941", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dbaf0211ef9de8b0b39a765fcef2c73bfb3e4b1c"}
{"id": "train_46942", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "reproducibility", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ff4ae34cfc8463732f4fd97b6130a3ce537f676e"}
{"id": "train_46943", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "tooling", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "54de4d6d318669347357ff35cc9bb08e6b74df20"}
{"id": "train_46944", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "documentation", "tests_are_truth"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "500053cea39fa7057c14f56c10cfbcdee84e6d98"}
{"id": "train_46945", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "auditability", "tests_are_truth", "governance"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1b5188e1c26b9beee427baaf79b5ae5d9c702b81"}
{"id": "train_46946", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "repo_scale_reasoning", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "113995afd2b8a4f0aec5c5c6b2a2fc0157d5323e"}
{"id": "train_46947", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "documentation", "security_gates", "reproducibility"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b1b0391d4849b7ea77be1e59a2930a9fd0980de8"}
{"id": "train_46948", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "reproducibility", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "244a3e9c8b7125d64500abc1aaeb7a77f73687de"}
{"id": "train_46949", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "repo_scale_reasoning", "cost_latency_tradeoffs", "governance"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "275aa9ed4857cda79904dd0ad02cb02d725bb106"}
{"id": "train_46950", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["governance", "auditability", "tests_are_truth", "reproducibility"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1e64cec33eb13e61bce3e09c0b3bee1cf72fe559"}
{"id": "train_46951", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "evaluation_metrics", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "abaf6989344bb05faa478e8deb962e21101f4ef3"}
{"id": "train_46952", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "governance", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2a19e47d8fc476101ce258a670ae20200286acbd"}
{"id": "train_46953", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["governance", "documentation", "reproducibility", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e34a100e64c6b2f8725e76edb255f082d211e2f6"}
{"id": "train_46954", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "documentation", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "08e4e9e36347f88c3b956ab6cd18440ef4b4d5f0"}
{"id": "train_46955", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "documentation", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c6d89f99e561ec5bd1d7462003f09ddd67045ad8"}
{"id": "train_46956", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "reproducibility", "tooling", "evaluation_metrics"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ca69bbebfe5cdf1a011f62eca392f4e3137b687d"}
{"id": "train_46957", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "governance", "repo_scale_reasoning", "tooling"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "223989e58503a388102d56aabcc78174266309de"}
{"id": "train_46958", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "reproducibility", "evaluation_metrics", "security_gates"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "049c22ba1c3e1bb22964a9cf55a21811efe4b911"}
{"id": "train_46959", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "documentation", "tooling", "ci_integration"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "636c2893ae2dc2efa777e79cc904ea28f54f3c33"}
{"id": "train_46960", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "ci_integration", "security_gates", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b9c742c2348a747d74d9ac588052ee8645a284a3"}
{"id": "train_46961", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "security_gates", "tests_are_truth", "documentation"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "70fffef3fe1e848984682ee96998bbfcc46e4539"}
{"id": "train_46962", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "documentation", "reproducibility", "tooling"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "16682bf51f065de16c820e56bb3fb09fd32ae4d8"}
{"id": "train_46963", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "documentation", "governance", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "81e4accd51a3f933800d5718b19fce28e42bc0cb"}
{"id": "train_46964", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "reproducibility", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "edcd5fe162b95ff34b5d2154b9e53581491c6817"}
{"id": "train_46965", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "security_gates", "tests_are_truth"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6d7af6cd57def1ca399e9da51c981ca02e630896"}
{"id": "train_46966", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "ci_integration", "tooling", "tests_are_truth"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1f15ab0e8f4db859209278ac672d2fafa985afde"}
{"id": "train_46967", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "auditability", "documentation", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d651ff4c0dc0eaf80178e370afb40fd4ef6dce14"}
{"id": "train_46968", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "tests_are_truth", "reproducibility"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "255b185fc9f3720c7e950f719c173740a940ce70"}
{"id": "train_46969", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a29983b15055410a9092f9267f34abe0ff2ab40b"}
{"id": "train_46970", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "cost_latency_tradeoffs", "documentation", "reproducibility"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2a8e55ca1ee9e0620b7b5d5a7b0ede45929f5030"}
{"id": "train_46971", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "documentation", "security_gates"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8b36e88c83bc5ad71ccb6d41a83cf5d143d2a0a1"}
{"id": "train_46972", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "tests_are_truth", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d57cdd8dce58598d4793fa4429a103c69da9c13a"}
{"id": "train_46973", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "documentation", "auditability"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ad6275c44033823773a2dd7f90a4a2ec5689a2cb"}
{"id": "train_46974", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "security_gates", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5fcf81e73fbd7a98b91077af94fbe4dc8f87c708"}
{"id": "train_46975", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "reproducibility", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4ab8126298dddebba3c77bbe140dc2d9c1ba2d87"}
{"id": "train_46976", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "repo_scale_reasoning", "evaluation_metrics", "tooling"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9eea7e74edbbfd109b09715c98bb566e6c686922"}
{"id": "train_46977", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "tests_are_truth", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b29853a339741678a579dd988f9d20c5a5f46a59"}
{"id": "train_46978", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "tooling", "security_gates", "documentation"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1155cf217b74f8a0a64b9561d763879209265c2f"}
{"id": "train_46979", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "governance", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b21f8e520d6eee25f41da6105f47a273a538c39a"}
{"id": "train_46980", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "repo_scale_reasoning", "documentation", "auditability"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "634d277472ca65d91cfe88e1245ea1c70bc351be"}
{"id": "train_46981", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "tests_are_truth", "documentation", "security_gates"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7857005340fb94de39f5ad336cf1cfe82d26ea37"}
{"id": "train_46982", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "reproducibility", "governance", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "805593c714e5bd5aad37c24fa93e5902b8d60413"}
{"id": "train_46983", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "auditability", "evaluation_metrics", "governance"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f4b58af8891e9f860120f66738ecbdeb3b3b30cd"}
{"id": "train_46984", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "documentation", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9fd5e5dddde41bcd8beba7fc4ca0796cab926ee5"}
{"id": "train_46985", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "governance", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ef23e48d9200c1482cd97a81070397870d6545c0"}
{"id": "train_46986", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "governance", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c846bc5b10ffaa843ffd7a6feeda79c4a15544f3"}
{"id": "train_46987", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3e84ffe4360b8c72447035fb4bdf77957e4cd6b3"}
{"id": "train_46988", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "tooling", "auditability", "governance"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ddd0712ec8bf14975db9ca24da83d47ce1595284"}
{"id": "train_46989", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "reproducibility", "tests_are_truth", "security_gates"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6061700c7b21f59048ae32530f6b0406b66322af"}
{"id": "train_46990", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "tests_are_truth", "security_gates"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f597bebe8e696583390bcc9e1a8159efe6ff14b8"}
{"id": "train_46991", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "documentation", "auditability", "tests_are_truth"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ffd57f40fc7fd6dd99b5af582295997674fccaba"}
{"id": "train_46992", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "security_gates", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "08cc082fd10f3d610300af03fa30df9b03b955c8"}
{"id": "train_46993", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "ci_integration", "documentation", "tests_are_truth"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9074539fa9a0aee5978d8bca9b55362e4841c3aa"}
{"id": "train_46994", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "documentation", "security_gates", "reproducibility"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "59b865ba76dbb2ae0a891fa7ad1da3c76a7ead78"}
{"id": "train_46995", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "tooling", "evaluation_metrics", "documentation"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1992fd9829eb2d9fcf7c6e3ae9c35cd7499bf512"}
{"id": "train_46996", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "reproducibility", "tests_are_truth", "ci_integration"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b70070f3a6d79cd11b5b55e60f9d7dcc2ecff487"}
{"id": "train_46997", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "evaluation_metrics", "auditability"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c34f2f3bab8bee6d519f8e5afb3c076b18253fc3"}
{"id": "train_46998", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "governance", "repo_scale_reasoning", "documentation"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d06dba29cc434110560eb6549ec834777af299af"}
{"id": "train_46999", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "auditability", "tooling", "repo_scale_reasoning"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "98989878f5b841684f9c64f0e7f83640534a9a8f"}
{"id": "train_47000", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "tests_are_truth", "documentation", "evaluation_metrics"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e3cc2a55268e8bcd9aced96c8854a7a83a4553b6"}
{"id": "train_47001", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "documentation", "governance", "tooling"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8a3064db4c28bc24e790795d0bbf5fd654e6bc4f"}
{"id": "train_47002", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "documentation", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6e7fc83b8ae5b2b23bbc2583f753879a6f57cc27"}
{"id": "train_47003", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "documentation", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a03204516e8fe96a3b5898b563ec47249c832311"}
{"id": "train_47004", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "ci_integration", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4aa82ff10dcf247f0e4e560e27f86dbe8bf636bb"}
{"id": "train_47005", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "security_gates", "ci_integration", "tooling"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fb2dd1b33d0872693d296e98a093ebbfeb401e73"}
{"id": "train_47006", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "repo_scale_reasoning", "auditability"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "00b8acd06343d5dc24ebb5651d4351cc57356166"}
{"id": "train_47007", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "auditability", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a9f4276a7da35e46d563af52c31eadd3452630e3"}
{"id": "train_47008", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "security_gates", "reproducibility", "documentation"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3ee908237be55374a408f95dc99e820436eb6798"}
{"id": "train_47009", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "tests_are_truth", "auditability", "documentation"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cc271e0c0815e61f4565057b6e2bf0e488c4ab20"}
{"id": "train_47010", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "tests_are_truth", "documentation", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ac679133818e23856adb5d0cdefb1b14952fc54d"}
{"id": "train_47011", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["governance", "documentation", "evaluation_metrics", "reproducibility"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0a25dd7c1ba82b1446d96d854c827dab1e7f4a04"}
{"id": "train_47012", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "evaluation_metrics", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e143d352b973ff15bf59858f068ffd0040501110"}
{"id": "train_47013", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "governance", "tooling", "tests_are_truth"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6a636194d614f2e89acb311a5b4876bc63ea3e92"}
{"id": "train_47014", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "evaluation_metrics", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "61be3fec5a75db53b7cc4a384cd400f0dd8e97fa"}
{"id": "train_47015", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "governance", "tests_are_truth", "reproducibility"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5998b0a376ee36093ed8552bddd786e47302e07f"}
{"id": "train_47016", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "repo_scale_reasoning", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a14d7ea6b226f3b8a18df7227ec1b2cdd7d842e5"}
{"id": "train_47017", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "reproducibility", "documentation"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6de0fcce71a404695696530b295adc17db0784e2"}
{"id": "train_47018", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "ci_integration", "reproducibility", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "900b16aed42d4bfe3f27c611a9757a2dafebd7e1"}
{"id": "train_47019", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "reproducibility", "documentation"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7ca4b2353e525d96f2b2043470a0a9e0db32d5f8"}
{"id": "train_47020", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "governance", "reproducibility", "tooling"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f47110e803bc94d1b3efe339d01e3d778dbaeaa8"}
{"id": "train_47021", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "security_gates", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ec1d17f881b0b9d0b664f2186399f5dffe23b21f"}
{"id": "train_47022", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "documentation", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ea7cdcefcc92ff9dd522a77ef6fa67505da6e373"}
{"id": "train_47023", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "tests_are_truth", "security_gates", "tooling"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9613d55b59f85767bbf5ff6465b7d90e869e20c0"}
{"id": "train_47024", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "reproducibility", "auditability", "tooling"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7f3357e04defac47f5ced30fd6a8184e0ff6265d"}
{"id": "train_47025", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "reproducibility", "documentation", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "32b8dcd01d3bf48a874d8cc332552d35b497b80d"}
{"id": "train_47026", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "auditability", "governance", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3768d71cd78f2e0cd11b4c8f85d513cf75b45017"}
{"id": "train_47027", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "documentation", "ci_integration", "auditability"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "28e7a8ece9f3d544db0b606a0a32ee94dd0ecb7a"}
{"id": "train_47028", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "tooling", "auditability", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ad49338b7f8ef8f42b778c64f44f3cf9ae62296d"}
{"id": "train_47029", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "tooling", "governance"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2ea8aef99d5e6e5a2e3f6fcb587b2a8205574c57"}
{"id": "train_47030", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "tests_are_truth", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bcad9d5f7885ecf03961aad75ec0e9a84301bac8"}
{"id": "train_47031", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "reproducibility", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2d9a278f7ae532c247286c0075f2da7af3c0f6c6"}
{"id": "train_47032", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "governance", "tooling", "ci_integration"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f94561b4bb3914f5365ad77d9e9936174ee46ce5"}
{"id": "train_47033", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "ci_integration", "governance", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8be5c1359144136dd30c366c53b30abfffe0a337"}
{"id": "train_47034", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "evaluation_metrics", "reproducibility", "tests_are_truth"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ab7ffacff5e2f307fa820066718a7a364f606d3c"}
{"id": "train_47035", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "reproducibility", "tooling"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "16c3c34678a6bd461758816d3555eb35a0db7c07"}
{"id": "train_47036", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "governance", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f5e7cf0667b14475c356ef48dba69a91ad6ef061"}
{"id": "train_47037", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "evaluation_metrics", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cedc8da500681f0546bc256ed481d8ca94fc8867"}
{"id": "train_47038", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "security_gates", "reproducibility", "documentation"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bb83e550ac2f16c7c1376f3e6072162d39720aec"}
{"id": "train_47039", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "repo_scale_reasoning", "reproducibility", "documentation"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ead81f19b99db50afb0299eb7f44bd850cf396ed"}
{"id": "train_47040", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "tests_are_truth", "documentation"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "02bbd351d685048a1e645d3f8dcc28e10cb8e18c"}
{"id": "train_47041", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "governance", "tooling"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1840ce37ca3e3272f654a5d51813868d9b4ba537"}
{"id": "train_47042", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0867cf4476cc7cb09528794770dc800665dbf536"}
{"id": "train_47043", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "documentation", "tooling", "tests_are_truth"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6bf35917f94b8e647014d60d2542a336651bf58f"}
{"id": "train_47044", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "tooling", "governance", "tests_are_truth"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2a51d7d9a4a0fdfccf72b90e7f085a9df7c7eed4"}
{"id": "train_47045", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "governance", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e899f432848221a42343c9e48e3ed4c77942012f"}
{"id": "train_47046", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "ci_integration", "tooling", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2be8cff0ecb2e6ff20385a8c3302d7f864503503"}
{"id": "train_47047", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "tooling", "ci_integration"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c84e941687c33a8647e709bd52112414e172db08"}
{"id": "train_47048", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "governance", "documentation"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "68916c151b27f29a0ff71bdcbf0afa8a62739a4d"}
{"id": "train_47049", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "evaluation_metrics", "reproducibility", "auditability"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "926b754563f02daf27756e13be3e0f8e4440f6d2"}
{"id": "train_47050", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "security_gates", "ci_integration", "evaluation_metrics"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d87844f09f7425293ecbf4b3d7332df049c4d1a8"}
{"id": "train_47051", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "evaluation_metrics", "tooling", "ci_integration"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "13b417479c35c839c75895eb9478b815b292ecd7"}
{"id": "train_47052", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "security_gates", "tooling"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d799612352a8c299ac591ebad19e10f9d2cbde17"}
{"id": "train_47053", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "ci_integration", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2cca50e8258a6c95434b8214df77395d44390b71"}
{"id": "train_47054", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "evaluation_metrics", "documentation"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ec59745b46af53a2fb26f8eb61180b7263363efe"}
{"id": "train_47055", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["governance", "tooling", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "06ca75b78dc06d9b72c74b8ebee215ea1661c420"}
{"id": "train_47056", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "repo_scale_reasoning", "auditability"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "54baab375159da8b6ebb50f6f375218e705b52ea"}
{"id": "train_47057", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "security_gates", "repo_scale_reasoning", "documentation"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "af2b8c3eb6353ad82adb36b01da55cc52bf6c05d"}
{"id": "train_47058", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "auditability", "tests_are_truth"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "831b82779e9a316e033b1c1697feeda035a70689"}
{"id": "train_47059", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "tooling", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc62b2e8b6593948ea77e4836298fac98176b4d1"}
{"id": "train_47060", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "security_gates", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7ee73b6b9082f08728f45d11ab84993bbb1a2057"}
{"id": "train_47061", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "cost_latency_tradeoffs", "documentation", "evaluation_metrics"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e0bec71f571d932cddd6f10940f15dd5577823da"}
{"id": "train_47062", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "reproducibility", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2ac998c08f734f2d5297aad1a3d9b4487393b7de"}
{"id": "train_47063", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "tooling", "security_gates", "evaluation_metrics"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6638157c33d110eb7241af4d01f8931af5385146"}
{"id": "train_47064", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "auditability", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "437de436df6ec0349d4e034cd7a560efe865282c"}
{"id": "train_47065", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "repo_scale_reasoning", "auditability", "tests_are_truth"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b335e901c2e5d0885d12d50ab5854f61b2202bdb"}
{"id": "train_47066", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "governance", "documentation", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0a3125e09162fc63021379bf73a33782ab960df4"}
{"id": "train_47067", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "cost_latency_tradeoffs", "security_gates", "documentation"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3312ab162ad54ea0323c2a071cd1afa20b0157e1"}
{"id": "train_47068", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "reproducibility", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3459d317f3d4dcbe71806cfc80fad42024325853"}
{"id": "train_47069", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "auditability", "ci_integration", "reproducibility"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "008730e7c37f5f670078db504ba882eb474e532d"}
{"id": "train_47070", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "security_gates", "governance", "auditability"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d89fabf4e57f1adf45d091a12be00c9e23d70812"}
{"id": "train_47071", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "security_gates", "auditability", "reproducibility"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "576f549bd241fdea166c8b91d68adef41d11c756"}
{"id": "train_47072", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "governance", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f3875f0f0aa729e265dbe052eaf28a236fa938ed"}
{"id": "train_47073", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "auditability", "governance", "reproducibility"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "af42c14d55bee9a3aad5c9b1e14d128809f686c9"}
{"id": "train_47074", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "tooling", "reproducibility", "governance"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "453a45c92525623c30c7425d377d7ea6d6605e91"}
{"id": "train_47075", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "reproducibility", "documentation", "auditability"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "966ec1b9fdb735c68be9b941f97f019f632e6417"}
{"id": "train_47076", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["governance", "reproducibility", "tests_are_truth", "documentation"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc1f3084772e63294533c4ad5c942c4c847bb1c3"}
{"id": "train_47077", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "tooling", "security_gates", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f1e12e80330b25794caedfcee08452fb87d50507"}
{"id": "train_47078", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "repo_scale_reasoning", "evaluation_metrics", "tooling"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "62162af32264eb20e045cd71aa7f34c1caa8b039"}
{"id": "train_47079", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "repo_scale_reasoning", "documentation", "security_gates"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1d8def0861a9907f767265c1076085995b79aeeb"}
{"id": "train_47080", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "documentation", "tooling", "governance"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "98905552fa64ed883f10dae5376ac86485af1fc0"}
{"id": "train_47081", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "reproducibility", "auditability"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6769bcb215ee73c3d4d46905f01510aa9abc695a"}
{"id": "train_47082", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "reproducibility", "evaluation_metrics", "governance"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1d34b60b04ac2f65d17935f058bfce4c39c5fbda"}
{"id": "train_47083", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["governance", "tests_are_truth", "tooling", "documentation"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8468fb4c4e7f31afc0a1a53b9ee0be57c1676bd8"}
{"id": "train_47084", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "ci_integration", "security_gates", "auditability"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "939a0d78adab12a0142a58801291239fdbe664c7"}
{"id": "train_47085", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "documentation", "governance", "tooling"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6f5b714b012331f6ba2c0093335158bcd2d8bc2a"}
{"id": "train_47086", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "documentation", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ee95e4657b38ec4334000acab6e4698a9483d348"}
{"id": "train_47087", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "auditability", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0ec6275b1f5b3ddd4e119ddc2570469d4db1481a"}
{"id": "train_47088", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "security_gates", "evaluation_metrics", "reproducibility"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "72516350faf61f3db0e10008aafd1063c654017f"}
{"id": "train_47089", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "documentation", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "00ae3374b69c8340abbe958a5cdb215dc9635a08"}
{"id": "train_47090", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "security_gates", "governance", "reproducibility"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1f0d1e9f39dcb75fefd39f65b1bf95b8bdcf8d84"}
{"id": "train_47091", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "reproducibility", "evaluation_metrics", "documentation"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aab70eba5378d8ea5cf8e87556a44fef351abd8a"}
{"id": "train_47092", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "auditability", "governance", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9b0dd3d0c11f4bbe2da50ed323870408a6fe68a6"}
{"id": "train_47093", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "security_gates", "auditability", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d354328221f1009db4f6cffa275aeefb8a1fdf77"}
{"id": "train_47094", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "reproducibility", "governance", "auditability"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ed096344c88ca9c5ab00dddf648338303c8c55e2"}
{"id": "train_47095", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "tooling", "reproducibility", "auditability"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ae2f1907e0c54b2dfd820b3034eb5bc349ce0d72"}
{"id": "train_47096", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "ci_integration", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "675b24de745a0b477bd163a17baf60e9f2d966a2"}
{"id": "train_47097", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "auditability", "documentation", "governance"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "47a45c3bbe827362d434b8f7b20406e1e74d2737"}
{"id": "train_47098", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "ci_integration", "reproducibility", "security_gates"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6781c8e633032869a91e3eeac7b86f6129e14136"}
{"id": "train_47099", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "ci_integration", "cost_latency_tradeoffs", "governance"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a8ffd31a63f06aa014ce2fb7bd91479fc8bff376"}
{"id": "train_47100", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "tooling", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "42294db117eb5ba876939dd3e8994b3275ee80d9"}
{"id": "train_47101", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "repo_scale_reasoning", "reproducibility", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4465e0e68ab0c1d3e49b9509684f03e638890a38"}
{"id": "train_47102", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "governance", "tooling"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d69c69e1da7edee86d64e6738b4f853ff9be0355"}
{"id": "train_47103", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "evaluation_metrics", "governance", "repo_scale_reasoning"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fb5acfab9afea48dd8a5eeb36c1e19de9a09c8d1"}
{"id": "train_47104", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "cost_latency_tradeoffs", "ci_integration", "security_gates"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9fc95a0c0bc83f0aaf5e017b144ef93275660274"}
{"id": "train_47105", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "tests_are_truth", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "94c0622c6066b746431c89dab0441f7ccba20d0a"}
{"id": "train_47106", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "ci_integration", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ad9abd7f58c33c3dd3db8fb10a1d0e8b04db01c8"}
{"id": "train_47107", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "governance", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e8a6d93424836645e00fc27e783de554431e3323"}
{"id": "train_47108", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "reproducibility", "documentation", "evaluation_metrics"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8460af8b218f9970a874de34db91ef5f9170264b"}
{"id": "train_47109", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["governance", "reproducibility", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "16d40bf08934fbc619e0cc0cc7995f3ad436e077"}
{"id": "train_47110", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "ci_integration", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8999a79cfbb6edb6ac392ff2c1f29f36ebdb855b"}
{"id": "train_47111", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "security_gates", "ci_integration", "auditability"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7da334b1b66e251ba71499617816e1aa23250533"}
{"id": "train_47112", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "repo_scale_reasoning", "tests_are_truth", "reproducibility"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7f0e6c1a0d58e2d5a5bed6a153ac6bcce0fd3c6d"}
{"id": "train_47113", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "governance", "tooling", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e1bf5e5de8cce6a311cb13ff2f2b2bcf57290447"}
{"id": "train_47114", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "security_gates", "cost_latency_tradeoffs", "governance"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ca96d955bda05f14802a96be31ddb1242a749de6"}
{"id": "train_47115", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "auditability", "ci_integration", "tooling"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8e17c46c2129e33cedbe3da27ee634144f7a3f58"}
{"id": "train_47116", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "governance", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cbe7a6d1379a5b8fcb3cf70b36c27d3a453205b0"}
{"id": "train_47117", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "documentation", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "13d95d7e424df5872bab9cd9076956841fb7031e"}
{"id": "train_47118", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "repo_scale_reasoning", "documentation", "governance"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d23731ada95401261bfe3eb77bf7a801471128ed"}
{"id": "train_47119", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "tests_are_truth", "evaluation_metrics", "documentation"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1b10951bfb8e8237aed7f8a1f2c417447375d298"}
{"id": "train_47120", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "governance", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "894ae7c212462cb0ce4d74137b19c2bc64408f28"}
{"id": "train_47121", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "repo_scale_reasoning", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "72bfb6f57bf2df0701e45b84e32a4f61f227348f"}
{"id": "train_47122", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "auditability", "tests_are_truth"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "777a20f4cf4e4b4c212d7cf116e3cf382258502a"}
{"id": "train_47123", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["governance", "ci_integration", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "be3ec1caefc322c60d37b5a0a0f9aefcfd447a91"}
{"id": "train_47124", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "repo_scale_reasoning", "documentation", "tests_are_truth"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3dc150d3cc083c71a1af60ae49e02ef31545ce22"}
{"id": "train_47125", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "tests_are_truth", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ae9997f129bdf8d9e499c86c40e5b2d0b3296e0b"}
{"id": "train_47126", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "documentation", "tooling", "ci_integration"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a37fd6b24d73382cfc4bc130e27095bcfc21df72"}
{"id": "train_47127", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "governance", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ccfb3cf813de1f1798fdc5a890219a7fd0c498b9"}
{"id": "train_47128", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "tooling", "reproducibility", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4cb3fd1a71ec7bc2a28d8a5c5fd527edae44cefd"}
{"id": "train_47129", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "reproducibility", "governance", "documentation"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b611c34d7509f23f6e68ad8548ad5b9c159da334"}
{"id": "train_47130", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "governance", "ci_integration", "tests_are_truth"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c3c3d5468f95791adf22119ba56669c4a140bccb"}
{"id": "train_47131", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "auditability", "documentation", "repo_scale_reasoning"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8b4fc65d69071d01247c50d3b41733a3cf144bd3"}
{"id": "train_47132", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "repo_scale_reasoning", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c0413a264900a05a879c131a72f642c359451d5f"}
{"id": "train_47133", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "cost_latency_tradeoffs", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "42b0f395687d54b1d49afcffa8f6e3583c61bac5"}
{"id": "train_47134", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "reproducibility", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fbb53c4faa101df3ff5af92df7b5369f1ce3ef34"}
{"id": "train_47135", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "reproducibility", "security_gates", "governance"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "979ebf0dd8b43a62d8bc4b402c6d6ebba64e598d"}
{"id": "train_47136", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "ci_integration", "repo_scale_reasoning", "auditability"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f97c030ae0cd957c3835677687ece12ed401d05d"}
{"id": "train_47137", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "security_gates", "governance", "reproducibility"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "767ac1923594a1f790bdc7cec8a45d5c39dab5aa"}
{"id": "train_47138", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "ci_integration", "evaluation_metrics", "tooling"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cc078cb656dd06558f56641bc0ab179ea0f9e735"}
{"id": "train_47139", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "reproducibility", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3cd192947b6c36be018fc4c47fff1ee42f1fedb7"}
{"id": "train_47140", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "auditability", "repo_scale_reasoning", "tooling"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ac9626d85d968ec6e0dc51a210b4d0d582269076"}
{"id": "train_47141", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["governance", "ci_integration", "security_gates", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ba4d169df3b4b49309b5773a30900be9286f1cb0"}
{"id": "train_47142", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "ci_integration", "evaluation_metrics"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7745199450edee0aef1440942c53d5c3e17d6bbb"}
{"id": "train_47143", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "auditability", "evaluation_metrics", "ci_integration"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "68f6644974c97ad36de86cbdecf814361808376e"}
{"id": "train_47144", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "cost_latency_tradeoffs", "security_gates", "evaluation_metrics"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b1498765568d5f6969778c72e9330f59be161a97"}
{"id": "train_47145", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "governance", "tooling", "documentation"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "78920b2b27bfe6fae6cffa2526204c02d186f0ac"}
{"id": "train_47146", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "governance", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "111baa041f307ad92fe00a2bd2c8ef58cc074b76"}
{"id": "train_47147", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "security_gates", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fad6e1caf392386fcaeddf1531e121b427046937"}
{"id": "train_47148", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "documentation", "ci_integration", "reproducibility"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5ee019972aca470f3a5c7b7fb22f70a93e7112a3"}
{"id": "train_47149", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "governance", "ci_integration", "tooling"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "527c305ccfc6a1890fa9dc6b7fbc09eb84d19590"}
{"id": "train_47150", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "governance", "auditability", "ci_integration"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3da39b6fbc03e3de3e49d0745809551f39b1c113"}
{"id": "train_47151", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "tooling", "governance"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d7649af5d4f27d370206d7cebe771cba562025ab"}
{"id": "train_47152", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["governance", "cost_latency_tradeoffs", "security_gates", "tooling"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7c34933fa9a53a08f4d4772498580e4fe8667274"}
{"id": "train_47153", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "reproducibility", "auditability"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d22e25bf4fb4624870fd285e0c9f7f19dbdc546d"}
{"id": "train_47154", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "ci_integration", "documentation", "tooling"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f429e4d8a10b45ca0a5912ff068eeb3907c1e907"}
{"id": "train_47155", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["governance", "tooling", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2ef461d1fb1d36e485f91fae5f6e7cfc7c069b1a"}
{"id": "train_47156", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "repo_scale_reasoning", "governance", "ci_integration"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "770446aef23deef5f4733ae323c7bcf48adacb42"}
{"id": "train_47157", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "auditability", "tooling"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e3147673bc89f032149880601e00246c691a4847"}
{"id": "train_47158", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "auditability", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e88c3da74c2605cfb666c9ca0a67cbca7ad08e57"}
{"id": "train_47159", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "ci_integration", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "995bcbae683d43387bdadaa99a2bfe234792d021"}
{"id": "train_47160", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "auditability", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f1c6b3fd8ea4b966673e13a5f9dbf45dd8f8f36b"}
{"id": "train_47161", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "auditability", "ci_integration", "documentation"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a122310c146114cabd6c37c5ef81f6c84b5ff0b3"}
{"id": "train_47162", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "ci_integration", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e30027a9304757d4ee801ffb4d5bab3e8fdd4b96"}
{"id": "train_47163", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "auditability", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "996b7b00e539edeb30fe1ed521f725939d5ca80f"}
{"id": "train_47164", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "tests_are_truth", "tooling"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d65634d6915c018e6b26a32d744e21c6a67b6037"}
{"id": "train_47165", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "ci_integration", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "089f19bb5310d569243c25a4483c4bf0b60aaee4"}
{"id": "train_47166", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "security_gates", "auditability", "documentation"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "823e0067ec3f303579eec8d68b1575461176a9df"}
{"id": "train_47167", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5372814ac33df70157b3abc1058e73b45ffae256"}
{"id": "train_47168", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "documentation", "governance"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b099243d1bf7df9e95b17776aed572736b66a7a4"}
{"id": "train_47169", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "tests_are_truth", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "efe9c136ab6a36ce5c3fb4bb2597c27327eb7974"}
{"id": "train_47170", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "security_gates", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f28437e797ddebe867b3d920a94dea8829b1fac0"}
{"id": "train_47171", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "reproducibility", "tooling"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "249a3ddbc0a15fc496d8ffa5c183dce55427454f"}
{"id": "train_47172", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "auditability", "security_gates", "tooling"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a447e31eafdceae7c2703de170665e37b25b0328"}
{"id": "train_47173", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "tooling", "repo_scale_reasoning", "security_gates"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4c178888eecf7dfe4c314afca4ef8f79d3fc533a"}
{"id": "train_47174", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "tests_are_truth", "repo_scale_reasoning", "governance"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9762443951d03bb05c2ce971e259b7d96bf59d05"}
{"id": "train_47175", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "tooling", "security_gates", "ci_integration"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9567939b06505b735380a5b142c923de3e949008"}
{"id": "train_47176", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "evaluation_metrics", "auditability", "ci_integration"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8010fe499f95803b204053552a4e46e51ceace6c"}
{"id": "train_47177", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "documentation", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8f64aacd3a2b4d97a6d58c2fa7178b842f7ee922"}
{"id": "train_47178", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "ci_integration", "reproducibility", "auditability"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7b185884742440f41f0017609fef9962bfe8d62e"}
{"id": "train_47179", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "cost_latency_tradeoffs", "documentation", "security_gates"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e8632f4c3f15680d7dc192e6ac7324cdbfb292dc"}
{"id": "train_47180", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "reproducibility", "governance"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ef4f0391a042e7f89bc0d669d9d3cf8907fb8b0d"}
{"id": "train_47181", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "governance", "security_gates", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3a6bab8c101b13f4e44cd2cde82d01ceead69df1"}
{"id": "train_47182", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["governance", "cost_latency_tradeoffs", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e011aed9cd7b7b9b55ea484032b6114e1b3f196e"}
{"id": "train_47183", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "evaluation_metrics", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5536d4d4942da67d7157304409c447ed682fac72"}
{"id": "train_47184", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "documentation", "tooling", "auditability"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3dc6f8ec496ecc533992d48da79d1d7af80aff56"}
{"id": "train_47185", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "security_gates", "reproducibility", "ci_integration"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ac765edb38930627533bc365a641f85cdc113606"}
{"id": "train_47186", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "documentation", "security_gates", "tooling"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7da334b1b66e251ba71499617816e1aa23250533"}
{"id": "train_47187", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "documentation", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "93a63a0cc2e5140e9b5997253ccb26877cb61f0a"}
{"id": "train_47188", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "evaluation_metrics", "governance", "auditability"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b028537734256931af141a2a540918edea3021f9"}
{"id": "train_47189", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "tooling", "security_gates", "governance"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8d2b3e4b4ffa9c07f7af81ee9bd82730b08cd978"}
{"id": "train_47190", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "documentation", "tests_are_truth", "tooling"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d2a7b951d35258938ea61e251d28cb8e02a6fa2e"}
{"id": "train_47191", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "cost_latency_tradeoffs", "repo_scale_reasoning", "auditability"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2047ff50a8242174fb5fe00b95df99a99e820dba"}
{"id": "train_47192", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "reproducibility", "governance", "tests_are_truth"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b9083abfc0193653fa9002c3b0e8599d9b85f815"}
{"id": "train_47193", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "evaluation_metrics", "reproducibility", "governance"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f18ce8adf3f3a870f41706dc054399bcd6e89626"}
{"id": "train_47194", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "cost_latency_tradeoffs", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "59aa3643f7d862cacef70dec6d387ba511b3de32"}
{"id": "train_47195", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "security_gates", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5014eb56809eb114706ce5443909f6101d7ad443"}
{"id": "train_47196", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "repo_scale_reasoning", "governance", "ci_integration"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6ae225aae0f89f62b61f00840d5b835f4f46fd8e"}
{"id": "train_47197", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "evaluation_metrics", "reproducibility", "tests_are_truth"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a86da61187b5c0caab8701754a357ca827a782cc"}
{"id": "train_47198", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "auditability", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2c0cb02d2979ef213dd4ad360239e4eb79801b74"}
{"id": "train_47199", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "auditability", "repo_scale_reasoning", "documentation"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9cec28592b5c864824d519c25c889a74ba9e3d8a"}
{"id": "train_47200", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "security_gates", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "146bdac6e0330e567714108491322f0e0a5b44a2"}
{"id": "train_47201", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "auditability", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a6ed3ddca2981feb0a13bfe7311ad3ef1ab40e8d"}
{"id": "train_47202", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "ci_integration", "tests_are_truth", "governance"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a59f1862eb73d24a4cd42816884a590890866802"}
{"id": "train_47203", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "auditability", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5c1cf5c62403efd0b6ce70bd8de3ccea3417768e"}
{"id": "train_47204", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "ci_integration", "evaluation_metrics", "documentation"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f9e6dd372d6f303439310902eb2a9855a31c3644"}
{"id": "train_47205", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "tests_are_truth", "evaluation_metrics", "tooling"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5781997bf00a71a09a40e51836495025f185a9d6"}
{"id": "train_47206", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "tooling", "auditability", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1ce6641fe892c3f9b874c5e34151eaff4861b49a"}
{"id": "train_47207", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "auditability", "governance", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cb5e6780c46494185544c2cafd624e573e7114ce"}
{"id": "train_47208", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "evaluation_metrics", "documentation", "reproducibility"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0c32e3edc89299f42d6e41854d5be2e6f3ceb9a3"}
{"id": "train_47209", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "tests_are_truth", "documentation", "security_gates"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d055e06cb645861cd981b04e0f8c5480270ac1e1"}
{"id": "train_47210", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "tooling", "reproducibility", "security_gates"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ae6e829973e052a5776f6a0c11c6ed7aef5d70df"}
{"id": "train_47211", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "tests_are_truth", "governance", "auditability"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cfa1176d12372a9da1a496f934d2c0d327545410"}
{"id": "train_47212", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "auditability", "documentation", "evaluation_metrics"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc7d2ca2f5b1b2e456eed1fff7521a430d6f2595"}
{"id": "train_47213", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "ci_integration", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1eaa97f759118523b7895734f6b2e57fce939375"}
{"id": "train_47214", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "reproducibility", "security_gates"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4fbd732ed9a6805c754f6551e49aff347d2e4f27"}
{"id": "train_47215", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "documentation", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3371a530bdecf067d5c38f542b83257f04a0d339"}
{"id": "train_47216", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "tests_are_truth", "ci_integration", "documentation"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6caba9bc8818427013d920d8a5a0998d9cc04bc2"}
{"id": "train_47217", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "security_gates", "auditability", "reproducibility"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c597533f7908ee0e338913d2c6ea52dedb85e68a"}
{"id": "train_47218", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "evaluation_metrics", "documentation", "auditability"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a90c68956a153c7b15089ef1e5bd03735e19ba13"}
{"id": "train_47219", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "cost_latency_tradeoffs", "documentation", "reproducibility"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8304a18d128b25f724c3b006113574afa08fb066"}
{"id": "train_47220", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "evaluation_metrics", "auditability", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1c3609463f79d32dc5c68fc39a16e377c3119b28"}
{"id": "train_47221", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "security_gates", "reproducibility", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "de4940c0b85c80dee22b0e1acc468c1012dc1236"}
{"id": "train_47222", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "reproducibility", "security_gates", "auditability"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c464d2317e6f0785ff9ee15cec3ca2577ae8a498"}
{"id": "train_47223", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "repo_scale_reasoning", "ci_integration", "tooling"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ac269f005962e9fedea0e81fc96076e204cb9626"}
{"id": "train_47224", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "auditability", "evaluation_metrics", "reproducibility"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "92a37f59bbaa9cffee440d96c6712ce44a2768e0"}
{"id": "train_47225", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "tests_are_truth", "repo_scale_reasoning", "documentation"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "147b122748e928d7ec24c864c5225673778d6b58"}
{"id": "train_47226", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "governance", "tests_are_truth", "auditability"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b0ad47f24e8dbfc6a8c10726d7ffa97887202f61"}
{"id": "train_47227", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "auditability", "tests_are_truth"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e247ee10896c2e74ba4030faca2f5b84dd40486b"}
{"id": "train_47228", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "reproducibility", "governance", "ci_integration"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "75e4791051ae1edc838f41cd3ee6eff546cdc583"}
{"id": "train_47229", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "ci_integration", "documentation", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c16744573a7fe7bd66fb438c833dbc220fd73a82"}
{"id": "train_47230", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "repo_scale_reasoning", "governance"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a3feb1389744c37b074c32fc7d487f0a853081a7"}
{"id": "train_47231", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "ci_integration", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bfe4a00a69e7ecbb5ca8af31abb285611b48378e"}
{"id": "train_47232", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "governance", "documentation", "tests_are_truth"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9bfd9901a6a907683df316f245af03b282cde432"}
{"id": "train_47233", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "repo_scale_reasoning", "evaluation_metrics", "tooling"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1c1d21189f613d29afd356f6495fd4fea1c37276"}
{"id": "train_47234", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "tests_are_truth", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "738d6062e4472cbd108d0ee1e4924b280d39472f"}
{"id": "train_47235", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "tooling", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cda0093a46aa332ddf54bfd6d4699919d6b0cc41"}
{"id": "train_47236", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "documentation", "tooling", "repo_scale_reasoning"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "febcea60a10345a7668d4f2f4b7181b77d9f167e"}
{"id": "train_47237", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "tests_are_truth", "reproducibility", "governance"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a611c4e77331e03c19af4f5f7dda191a2e862448"}
{"id": "train_47238", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "governance", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "30109127191da953cbceb64695507b2454483fa0"}
{"id": "train_47239", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "reproducibility", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "492c12e8a64e2de8ea4c56ced652485430c01f9f"}
{"id": "train_47240", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "auditability", "documentation", "tests_are_truth"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6f0484c5b96cfda55e152709e397162f7148a6d8"}
{"id": "train_47241", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["governance", "tooling", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dce97c02ef828285b9bbcf6ae289405179ea0585"}
{"id": "train_47242", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "auditability", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4e6ea78c2570223264bf03d0c721d30443286e70"}
{"id": "train_47243", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "reproducibility", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0cc4a7f1604b0b5329fd07f1bb3b240759353ccd"}
{"id": "train_47244", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "documentation", "auditability", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8acd471390ed7b6737c35e325b548ebe7e4c98f2"}
{"id": "train_47245", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "tests_are_truth", "governance", "documentation"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a77817cbffcceec5bcd46f0213b00a1bf8813f8a"}
{"id": "train_47246", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["governance", "cost_latency_tradeoffs", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7e2f00fe100aa6b1172a720d19d9322057a12019"}
{"id": "train_47247", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "tooling", "auditability", "security_gates"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f1a4f50faff10f12454b6adade0b9d39fdabe8ec"}
{"id": "train_47248", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "05a8f74fd7660813efe25145178302cfb7ddc30d"}
{"id": "train_47249", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "documentation", "auditability", "tests_are_truth"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4dfa6f26921698a392efd255f418b3146d1d2646"}
{"id": "train_47250", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "evaluation_metrics", "repo_scale_reasoning", "documentation"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f8bb7de103f35b9690a6e29a9a1560d177105bdd"}
{"id": "train_47251", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "reproducibility", "tests_are_truth", "documentation"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3047b49c844962f195f0dce89b5a7f3ff31be92a"}
{"id": "train_47252", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "tooling", "documentation", "security_gates"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8b2c15efea71dfbc5ea685a91ec0bebb71d98af8"}
{"id": "train_47253", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "evaluation_metrics", "ci_integration", "tooling"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "92bf4c646873946335879df07261eb1aedf799d7"}
{"id": "train_47254", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["governance", "cost_latency_tradeoffs", "ci_integration", "tests_are_truth"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4e2c6a2a4955a9451d2e7da86a15ced7ed626504"}
{"id": "train_47255", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "governance", "auditability", "tests_are_truth"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9d9a62b37c1c9f2ffa9d6a702ee7232ce74904d3"}
{"id": "train_47256", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "auditability", "tests_are_truth", "reproducibility"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "30ba4c62348ea09afd8044c7e4d975335c0f65fb"}
{"id": "train_47257", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "cost_latency_tradeoffs", "security_gates", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "410c5fa7777ab31d4e6d9b71907707e70aa2c6ef"}
{"id": "train_47258", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "governance", "auditability", "ci_integration"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "167d305b15ace9390053486f5f42fa428f26b748"}
{"id": "train_47259", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "governance", "reproducibility", "auditability"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d6fa74bd7fd277bfcaf0ed3235be1842e43396d5"}
{"id": "train_47260", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "reproducibility", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c1e56064ac7a8677ce42c17b873aacf31f311108"}
{"id": "train_47261", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "reproducibility", "evaluation_metrics", "tooling"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc664b73f7aecd8de472171ff2c2d704518a19a6"}
{"id": "train_47262", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "evaluation_metrics", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "77a3447c97a30af53d51cf551e3efe7eba01bdd3"}
{"id": "train_47263", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "security_gates", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "805593c714e5bd5aad37c24fa93e5902b8d60413"}
{"id": "train_47264", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "reproducibility", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fdd6be1dc9e5978b2ceb01e061aac67d39003e25"}
{"id": "train_47265", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8da6080d73f1970c545a5968b4f6663c3b09c8df"}
{"id": "train_47266", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "evaluation_metrics", "tests_are_truth", "reproducibility"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "500fb53c6ed71b33cd8a868ad94e5ada527684a5"}
{"id": "train_47267", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "repo_scale_reasoning", "tooling", "security_gates"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cbc14cd93f18b2e2034f15c7a301baf4f2864808"}
{"id": "train_47268", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "documentation", "tooling", "ci_integration"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ff07afe53d3efe383a4cbac7a729cbb62d70e73c"}
{"id": "train_47269", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "security_gates", "auditability"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0eedcebcb6133de89affeaa90df0fd3511e8a85d"}
{"id": "train_47270", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1c6ebe922cb6cd7da98b71d9663b2a8731e7af26"}
{"id": "train_47271", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "governance", "repo_scale_reasoning", "auditability"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "181654a103c17503eebf871c648637e5458feb7f"}
{"id": "train_47272", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "security_gates", "tooling", "reproducibility"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e6fdcd9dd4f7032b1e41a89f67987c0d175d7980"}
{"id": "train_47273", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "governance", "security_gates", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7d2f24dbc5f59788b75ca5b2cb0895a6eb885c01"}
{"id": "train_47274", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "ci_integration", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1d988b2cff16d08ff83d0cc115f05d9a8b339c1d"}
{"id": "train_47275", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "documentation", "governance"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "51b9b2afb1794e548be54b3920df58bc4b9787d5"}
{"id": "train_47276", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "reproducibility", "evaluation_metrics"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cf18e40907b3d78787ba2c6dbb7119e77ec8a3f6"}
{"id": "train_47277", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "repo_scale_reasoning", "security_gates", "governance"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ef286bd0132c036b54d269d62535d02b11471943"}
{"id": "train_47278", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "tooling", "auditability", "documentation"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e13e032488895f8926dd7b5c4dedbb28a4774dc4"}
{"id": "train_47279", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "tooling", "auditability", "ci_integration"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "66833561885674b1503ef773f048f8f15f067bc8"}
{"id": "train_47280", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["governance", "tooling", "auditability", "tests_are_truth"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9ab13c2383f69a91797618fd5f69db8aafebf154"}
{"id": "train_47281", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "governance", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6b6c4d4c3784382a65b2c08ceea510396ed3fbb4"}
{"id": "train_47282", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "reproducibility", "evaluation_metrics", "governance"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "74d2b0fd1e15b3ad4bacfe8e73f889cf0d82ca4d"}
{"id": "train_47283", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "tests_are_truth", "security_gates", "governance"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bd0677105d3c4793f4a64af142eb5a3558caabab"}
{"id": "train_47284", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "tests_are_truth", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "da7dbb2e4a1e69c7c1f1db3e2056fdc2603aeb4e"}
{"id": "train_47285", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "tooling", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dab13c4b26a1b99d0747a19cdf0943a59e974cae"}
{"id": "train_47286", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "governance", "ci_integration", "documentation"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "84603585ce645f45cafc952f9bacaaa31c2e7d4e"}
{"id": "train_47287", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "evaluation_metrics", "reproducibility", "ci_integration"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "459fa43b6486e1e149c796025265d889b8945fbf"}
{"id": "train_47288", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "evaluation_metrics", "tooling", "ci_integration"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6e16556698934544e510e0cb2af92176f3824632"}
{"id": "train_47289", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "governance", "evaluation_metrics", "ci_integration"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "25b3752fe34f428df7c1010316a48930947a6157"}
{"id": "train_47290", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "evaluation_metrics", "governance", "ci_integration"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "215b0a8e0393adbbda41529b4f09ce711862686a"}
{"id": "train_47291", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "auditability", "security_gates", "tooling"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "212fd619d9dc946a5baf698982795a332c3523de"}
{"id": "train_47292", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "ci_integration", "auditability", "documentation"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3032205c6def0d1b97243fda49547ac462997167"}
{"id": "train_47293", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "ci_integration", "reproducibility", "evaluation_metrics"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dfc0ddb1fa04aa97fca19c3677007f627b1cb04d"}
{"id": "train_47294", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "security_gates", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "78da504c7d0c219674bb80e114cbdfb3919b6575"}
{"id": "train_47295", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "tooling", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "49c7b2b73ab0073fe8c9e6d046a166a2bdf00d60"}
{"id": "train_47296", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "ci_integration", "auditability", "governance"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c02e5e5a62b061d5a9d530bfec8b3c40fbec950e"}
{"id": "train_47297", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "evaluation_metrics", "documentation", "ci_integration"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0646cea336de9459a3a4ee16cec0a63fad745ff8"}
{"id": "train_47298", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "security_gates", "governance", "tooling"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "72f0f0b8c04bce5b6dc09c2eebd0d6a187d48a8a"}
{"id": "train_47299", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "ci_integration", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "447763d891167bc5694f6852b4066dd11cd5e4ce"}
{"id": "train_47300", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "governance", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3ed9c91c1ace68e99fa4de30bd444cab9309aca1"}
{"id": "train_47301", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "tooling", "auditability", "governance"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3682cb717e97be48e7853a32f4d93920e62015aa"}
{"id": "train_47302", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "tests_are_truth", "governance", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fc7d69779cb5feb55fd9481300673e43ab7a19c2"}
{"id": "train_47303", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "tooling", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5fd55962ba6db5fb769d4369da27cfd2687de97d"}
{"id": "train_47304", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "auditability", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "40174e19687055e4c49c9afbdc75b726d1c69e71"}
{"id": "train_47305", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "tooling", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "140e86b4092dafc1871811fe0763056d5b8ce605"}
{"id": "train_47306", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "documentation", "security_gates", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aa112931989d999693fe2bd0005a2234a7460cca"}
{"id": "train_47307", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "repo_scale_reasoning", "documentation", "tooling"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2be88a2f9a411c24e21c9b8e392699eabfef58d3"}
{"id": "train_47308", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "tooling", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e666cc1dafe2d9fea82d0b0c799109ee12d67c15"}
{"id": "train_47309", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "auditability", "documentation", "ci_integration"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "19f6f6da5c9b693657464ca7c00a781e67c576bb"}
{"id": "train_47310", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "auditability", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2a4152ae154a2c1cd83b0df249e8d91984f1704d"}
{"id": "train_47311", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["governance", "tooling", "reproducibility", "security_gates"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7e606c160449d50aa6c14c51f64d2998341df8f7"}
{"id": "train_47312", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "documentation", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a2d9d97ba66139ac53d7040343408a0e9992efcc"}
{"id": "train_47313", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "tooling", "ci_integration", "tests_are_truth"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "665099c93d14093c7e31f869b11f8ed6583430d3"}
{"id": "train_47314", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "tooling", "security_gates", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3770e38e6f7bcfa8c1d3a25722fdc4e02c2f78c2"}
{"id": "train_47315", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "auditability", "repo_scale_reasoning", "tooling"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "79ede69ca9491332e393f7596172cce91984d246"}
{"id": "train_47316", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "governance", "auditability", "reproducibility"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d5f4bc994779225cb34a9a6b3757769a6530a13d"}
{"id": "train_47317", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "documentation", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a21e406d6e5741ccf08b709070207f2108d66bdd"}
{"id": "train_47318", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "auditability", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "25aa2a7436a8f7c74c89fa201cd1bcf3fc7c2006"}
{"id": "train_47319", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "auditability", "governance", "security_gates"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8d99606b0cb9adf70735f0f29f668326f324febf"}
{"id": "train_47320", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "tests_are_truth", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ea11daaec00571dd191b656a1b4555ce1782d2f8"}
{"id": "train_47321", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["governance", "auditability", "repo_scale_reasoning", "security_gates"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc5b30ac9e03bb5953c096b2df0774e83794d038"}
{"id": "train_47322", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "tooling", "reproducibility", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "58c384731f7f15cbd15c03c17a612293ded0755f"}
{"id": "train_47323", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "auditability", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8e9fc70edb7b7ad72701071a30d796221dc9d443"}
{"id": "train_47324", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "reproducibility", "evaluation_metrics", "ci_integration"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "160f14673dd3353ede01c043efadc978c2f625fb"}
{"id": "train_47325", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "reproducibility", "documentation"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3e4dfd4814d2fa16f9f6b81fc5bc4dd52023012f"}
{"id": "train_47326", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "auditability", "governance", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0e001aed2dc97cf50ac4d102f9ce5015b14e835c"}
{"id": "train_47327", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "reproducibility", "governance", "security_gates"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3597c7ca99c84cdb9082a51e386314f8b1f6e884"}
{"id": "train_47328", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "auditability", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "08944b177c53ca133570133b1f7ee09b417b927a"}
{"id": "train_47329", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "evaluation_metrics", "governance", "tooling"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d0d1500829c49eabf8236be6ec005cfadee979a3"}
{"id": "train_47330", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["governance", "reproducibility", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b9f11264d9a489871d909b70182a044d70f9de00"}
{"id": "train_47331", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "governance", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c7dfb1d35e1705f57457f1a856467f2a0eafe202"}
{"id": "train_47332", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["governance", "tests_are_truth", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5bbab08a6958b7b73fe9bc1090f6610b9d6b3b41"}
{"id": "train_47333", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "governance", "evaluation_metrics", "security_gates"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2c970f5ebe3cced35d352a0f8c83a98492e01e0f"}
{"id": "train_47334", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "tooling", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1fe801f4f43d064bd05979f31fac54e906625d76"}
{"id": "train_47335", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "auditability", "documentation", "governance"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "12cfa5859ea5797be010fd811b21d70dda3c962f"}
{"id": "train_47336", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "tests_are_truth", "documentation", "security_gates"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "38dd67098826706fe61defb6e0d388d7c9014198"}
{"id": "train_47337", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "governance", "auditability", "repo_scale_reasoning"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3a628556eded50f88c9d445f5ab70d7e239143e3"}
{"id": "train_47338", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "documentation", "reproducibility", "governance"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2dc4d92e95c2b3c05ab2738afd337683d9580226"}
{"id": "train_47339", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "governance", "tests_are_truth", "auditability"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "007393e5fbefd49553579a35a64ec62c5c233eef"}
{"id": "train_47340", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "documentation", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "06356be88032a7c0cbcd1033438c4efec54b82bf"}
{"id": "train_47341", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "ci_integration", "tooling"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a68f948b9301ace61336c153a9ad16be9e33ba6d"}
{"id": "train_47342", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "tooling", "ci_integration", "security_gates"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "071296a5a3f33ed0f51600caaee96f1799d19274"}
{"id": "train_47343", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "evaluation_metrics", "ci_integration", "security_gates"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "96a113ecbde768764e98368168c8ae80dd7366ef"}
{"id": "train_47344", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "reproducibility", "documentation", "ci_integration"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1087713866939887d947f06be605f1b69be33bbb"}
{"id": "train_47345", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "cost_latency_tradeoffs", "governance", "reproducibility"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8b4a1c971aeb366ae37ea1e051b9cb1bf22478d8"}
{"id": "train_47346", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "cost_latency_tradeoffs", "evaluation_metrics", "ci_integration"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9a394575e27f3a40c4ad08bc997560feb6d92ffd"}
{"id": "train_47347", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "auditability", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "84dc69c88b1b0d29131081ba7559b36fee02e3ef"}
{"id": "train_47348", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "tests_are_truth", "ci_integration", "documentation"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "65141b34825320c6fe087557d9508f9859c320ce"}
{"id": "train_47349", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "documentation", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2464f7265b27cd7d9174575eb44d07525ebd5e2f"}
{"id": "train_47350", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "tooling", "security_gates", "reproducibility"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b5ddaa6c5fc78a228cc03a412c43a66b4ec943e3"}
{"id": "train_47351", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "security_gates", "documentation", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3316b30b10750cdff4d1412a061833baac29f78a"}
{"id": "train_47352", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "governance", "security_gates"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2f4b216579469f23165f844b93a1ba36c894890e"}
{"id": "train_47353", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "documentation", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e696eff0d09b9028b29e9edb218a3a005b915310"}
{"id": "train_47354", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "reproducibility", "governance", "documentation"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0b02f779d3922a9d384cf7c8d6b89983a90ee316"}
{"id": "train_47355", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "documentation", "reproducibility"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a0bba5703e43c7dba01893587c84a2d651e392f3"}
{"id": "train_47356", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "repo_scale_reasoning", "ci_integration", "documentation"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5fc5e253f2a16b3892b076727fd2b2b567da4eaf"}
{"id": "train_47357", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "reproducibility", "ci_integration", "governance"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6f5f548af1645b5d6ca35ace496704f3cdfbec90"}
{"id": "train_47358", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "tests_are_truth", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1dbdee4368083d91cf2f23a6642379d5cf3f6133"}
{"id": "train_47359", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "auditability", "governance"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "402a63a5cef26541bcdada3d743711945920f100"}
{"id": "train_47360", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "tests_are_truth", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "75c074f404d0981f39eea5fc923f0d6ea3c53a64"}
{"id": "train_47361", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "documentation", "auditability", "governance"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b3993a734a39be74ba0f600c7601aed0dcc3e432"}
{"id": "train_47362", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "auditability", "documentation", "ci_integration"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8f8809746638bd966a5daadc660f369e25fa115a"}
{"id": "train_47363", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "governance", "tests_are_truth", "ci_integration"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6f6359dfd71813231a1369c61f0af31b2b9aa991"}
{"id": "train_47364", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "tooling", "auditability", "ci_integration"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9eb4fd6514873538b5f1ca8ea4005096a20c1018"}
{"id": "train_47365", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "tooling", "evaluation_metrics", "reproducibility"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8267abd0e65ba53656820a186f7b5bb14539215e"}
{"id": "train_47366", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "tests_are_truth", "evaluation_metrics", "reproducibility"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8a8ddbf54da067197e9215a666babd6cf4c54b60"}
{"id": "train_47367", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "documentation", "tooling", "reproducibility"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3b6c5e88f388015daa1b3dcc106493b7e863dc90"}
{"id": "train_47368", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "ci_integration", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9738f03f5fe704480e854344ffb1b0d087041feb"}
{"id": "train_47369", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "auditability", "reproducibility", "documentation"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1e5d7e046d09d9f318d38bc420859b3a8a5bcc62"}
{"id": "train_47370", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "tooling", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bf22adfe0aa4c813868704dbd9bfad4d8424610c"}
{"id": "train_47371", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "documentation", "security_gates", "auditability"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4ed6169c9cdd118daf8da64d47d46cfa8b7f13d6"}
{"id": "train_47372", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "reproducibility", "auditability", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c592d178d54f58e2d00381a7385094df7e101209"}
{"id": "train_47373", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "tests_are_truth", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "af46df9a16004c19fe3deb42329d62b19c84ab8f"}
{"id": "train_47374", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "cost_latency_tradeoffs", "reproducibility", "governance"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "36d211c05ae340cd917d751c7c8603f4ace1f621"}
{"id": "train_47375", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "auditability", "documentation"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "363c4878f55a1912fcfdfa7e432f4b295f2b6c4a"}
{"id": "train_47376", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "governance", "tests_are_truth"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f2b0a8cdf2b1d1e5c7eb04bcc0551c89602067f1"}
{"id": "train_47377", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "evaluation_metrics", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "047690d9feff67c05602c7e7546441bd843ff0ba"}
{"id": "train_47378", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "security_gates", "governance", "evaluation_metrics"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b095922fbf5166761c30e1ff9c54ff72fbd33e71"}
{"id": "train_47379", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "governance", "ci_integration", "security_gates"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e86f35cbb2b5cd0103f8776561a8285263b14e7b"}
{"id": "train_47380", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "ci_integration", "security_gates", "documentation"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e61c0c4f365f05bcc07ac6ee974514dfaeee4eab"}
{"id": "train_47381", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "documentation", "auditability", "repo_scale_reasoning"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "faccf6e418051516c67a412b2bb85afbc7cd3ce4"}
{"id": "train_47382", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "ci_integration", "repo_scale_reasoning", "auditability"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4ca97f32a28d59180d6fb4d41489962ad5a5192c"}
{"id": "train_47383", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "documentation", "auditability", "security_gates"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e241e4e54420704eff078523b7862d081abf6e7b"}
{"id": "train_47384", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "tests_are_truth", "tooling", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ba57200bcd1261b0a4c984f0aa18489642c4cae7"}
{"id": "train_47385", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "documentation", "governance"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ddc01bdff018057adc844470db89bbf4307d3864"}
{"id": "train_47386", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "tests_are_truth", "governance", "reproducibility"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "83eebc8ef4b6bba4231e92b7d959590e2c514a30"}
{"id": "train_47387", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "tests_are_truth", "auditability"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e9520300739fc841022f8c585eb95e24997d179e"}
{"id": "train_47388", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "documentation", "cost_latency_tradeoffs", "governance"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "58e36ef314bc7d3d57365693253ec3cb49349a1e"}
{"id": "train_47389", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "repo_scale_reasoning", "evaluation_metrics", "ci_integration"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d0c69f56440eb20237ccf481ad4c289d43edb734"}
{"id": "train_47390", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "tooling", "security_gates"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "53dbcb022f3afa261948e632f2d3c6eba2e8f750"}
{"id": "train_47391", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "tooling", "security_gates", "ci_integration"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7e0eadbbb6b91bb69fe663ca569190f1b6e8cd1d"}
{"id": "train_47392", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "repo_scale_reasoning", "evaluation_metrics", "documentation"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e66d24de0d2ee733c85aa639885eb37bfabe1f79"}
{"id": "train_47393", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "ci_integration", "tooling"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "35bf0bebae0d019e7c864519ea7666d5cdaa3d80"}
{"id": "train_47394", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "auditability", "tests_are_truth", "tooling"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8da14a0007ee6aabeccf594ab69e800b338a2978"}
{"id": "train_47395", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "56847a90fe8b0ff8b828121106d97092690082d7"}
{"id": "train_47396", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "ci_integration", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b030f28558c2952e41394c69e1a3a2939ba1adfd"}
{"id": "train_47397", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "ci_integration", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "747b351144caf5cdf363f03b7e05315ef2f75c15"}
{"id": "train_47398", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "tests_are_truth", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "548a8b0ee02b5f05a83bf271a2b4d19137e39970"}
{"id": "train_47399", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "auditability", "ci_integration"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a249d1ba334df64e161effb3f9e6afeebd1bca76"}
{"id": "train_47400", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "auditability", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8ae01491838a155f91980c42a857af540d9e840b"}
{"id": "train_47401", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "reproducibility", "repo_scale_reasoning", "documentation"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eab4c897ebf1d0cdd2a740e2be457a242b30c8d6"}
{"id": "train_47402", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "cost_latency_tradeoffs", "security_gates", "auditability"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "37de0f5518f02b8c861bf77e9eaf0f0889848a6f"}
{"id": "train_47403", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "ci_integration", "auditability", "reproducibility"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1f60b6dc4250fb54429e3cc447ca96d58cd2cd9a"}
{"id": "train_47404", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "repo_scale_reasoning", "tooling", "reproducibility"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "769882db81fd6b4077cb55062ce4e6586600fb4c"}
{"id": "train_47405", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "security_gates", "governance", "tests_are_truth"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7f7effbbe01eed39cbee8cc4ea4e3e4a6e9c973c"}
{"id": "train_47406", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "tests_are_truth", "ci_integration", "security_gates"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "350441f8d39b59d5b8262ade3615a95b783ec0c2"}
{"id": "train_47407", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "security_gates", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6a5322f219a06777b4e2bb1fc0bc8a3fda63a811"}
{"id": "train_47408", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "auditability", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c3c9ca18a6e87157c0673ff815ff7ae9990b9e62"}
{"id": "train_47409", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "ci_integration", "governance", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3394839e590444722c70a774e6088513f837a1a0"}
{"id": "train_47410", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "evaluation_metrics", "tooling", "documentation"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "889a51edcb34e0656885ea0451eec0f6343c0b42"}
{"id": "train_47411", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "tooling", "reproducibility", "ci_integration"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f8bb7de103f35b9690a6e29a9a1560d177105bdd"}
{"id": "train_47412", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "ci_integration", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "90a31dbe3f1451114046b5939921aaa595b2ae34"}
{"id": "train_47413", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "documentation", "security_gates", "repo_scale_reasoning"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "81244eca2219413bb2a44c460d2781caa7f53eeb"}
{"id": "train_47414", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "tests_are_truth", "tooling", "documentation"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f27bc74bec535b1318c77c05bde5e757cef0cdd2"}
{"id": "train_47415", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "tooling", "documentation", "governance"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b699c470304cb69e9566f3f8ed9b54a42a8ebcc5"}
{"id": "train_47416", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "tooling", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "273d2cd0398ad3ad37a67b9ab9b96511b1a45e40"}
{"id": "train_47417", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "repo_scale_reasoning", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bd93d1aad9990a6583ad9c6fa2e3957d024b9d16"}
{"id": "train_47418", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "tests_are_truth", "reproducibility", "tooling"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9c535316cc2d413424192ae85ac3b1f0bb51ca93"}
{"id": "train_47419", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "cost_latency_tradeoffs", "security_gates", "evaluation_metrics"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cf6003dff70abf96796778c9a6724cf3a6e4baa7"}
{"id": "train_47420", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "reproducibility", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f4bb26ca493bd48a313b02edc47c393fac6d0018"}
{"id": "train_47421", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "tooling", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c69ec497969a3037ba795ac827e5d4c1fd389db7"}
{"id": "train_47422", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "documentation", "auditability", "reproducibility"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b5a5492b31db2b43cababd37fdb4bd2a0c0e1ac9"}
{"id": "train_47423", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "tooling", "tests_are_truth", "reproducibility"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fa83eda8e9b9c3a7029903cb9ea2294b76310c58"}
{"id": "train_47424", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "auditability", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e899f432848221a42343c9e48e3ed4c77942012f"}
{"id": "train_47425", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "ci_integration", "governance", "tests_are_truth"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3643460dab1d7523b7ee1cf82c7e5df34c98ec1d"}
{"id": "train_47426", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "ci_integration", "governance"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5b304abf5822694d96f454fd48d7fc490e64ec98"}
{"id": "train_47427", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["governance", "auditability", "ci_integration", "tooling"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9422a0bd579a981d876f568aae15a6e5700ed1ce"}
{"id": "train_47428", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "auditability", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cb2a9301b8ac14b50ea16fbf2a0837c6f65998b9"}
{"id": "train_47429", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "auditability", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4abbe0f10f3302205b3fb64452a7731b4a7dc9a3"}
{"id": "train_47430", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "cost_latency_tradeoffs", "documentation", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "acba34185ae9e2a95aaddc1f2e8a7b4db9058edf"}
{"id": "train_47431", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "evaluation_metrics", "governance", "reproducibility"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cdc173b23a1b83f706773cfe3a9dfe82d9684ec5"}
{"id": "train_47432", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "auditability", "tests_are_truth", "reproducibility"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2ac43057c2e2da09bd5c79a07de653aa480952f3"}
{"id": "train_47433", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "documentation", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0a6a931fffd7071107bf73e33b92e4e73d9df778"}
{"id": "train_47434", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "reproducibility", "evaluation_metrics", "documentation"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "336041202edf3b12d9dddb92540fdca7fe1d751c"}
{"id": "train_47435", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "tooling", "governance", "documentation"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2a3e5bffadcc9865c1ee96c243ba4752cf7dc10b"}
{"id": "train_47436", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "tooling", "repo_scale_reasoning", "auditability"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "01dee36d3b30a4b4a8b4574f3d96cf839f3f82c5"}
{"id": "train_47437", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "reproducibility", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cdee0bfc761cbdc08513d0ece086225f1a50c734"}
{"id": "train_47438", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "tooling", "security_gates"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5cb906f5bd7e403094989d058d58b1e956381df2"}
{"id": "train_47439", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "governance", "documentation", "evaluation_metrics"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "57d5bea64ae3ddd318ac3d296fd45768c782715b"}
{"id": "train_47440", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "ci_integration", "documentation"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d5f03c08347bfeb090b8826576f1c9004b75ee7b"}
{"id": "train_47441", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "tooling", "security_gates"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b812c3d737bad321f966ed711633e70ea8a61fcd"}
{"id": "train_47442", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "governance", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "54d3f7e391a4de46fe83c45d3fa8f9f91edeb921"}
{"id": "train_47443", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "reproducibility", "governance"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d595907e9f47bb399886b063b300be21a841cfb1"}
{"id": "train_47444", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "evaluation_metrics", "tests_are_truth", "governance"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2fd0518d5399114fa0885ffdfa27bff2d663307c"}
{"id": "train_47445", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "reproducibility", "security_gates", "evaluation_metrics"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c5afbd6379237ed65e9bfa1afe0c44c269e2b9de"}
{"id": "train_47446", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "tests_are_truth", "auditability", "security_gates"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5f20566c3305ca425af1a11e387934afe25a646b"}
{"id": "train_47447", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "repo_scale_reasoning", "tooling"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "31c4bbe5faecf9f3853652bab042922cf3f36160"}
{"id": "train_47448", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "ci_integration", "security_gates", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3468f8f0aaa8978e73a9d3db2fc2303ef6221290"}
{"id": "train_47449", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "tooling", "ci_integration", "governance"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "086fb378f84a67d5449906135e62835ae3089dbd"}
{"id": "train_47450", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "reproducibility", "tests_are_truth"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5cfa908d17cf4eb3f5c3b862eacd56dd10729fd8"}
{"id": "train_47451", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "evaluation_metrics", "reproducibility", "security_gates"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "35fb19826689699d02a0cac20960ace1682d2040"}
{"id": "train_47452", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "documentation", "auditability", "tests_are_truth"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aab25c26ad9f6b8e2f9b7b56c460fffc5c584d4e"}
{"id": "train_47453", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "documentation", "cost_latency_tradeoffs", "governance"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b0707fa06fa358d37192996ac46eba396e642122"}
{"id": "train_47454", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "evaluation_metrics", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "55f2e2001f55b2ff56874413708b8dad01d7312b"}
{"id": "train_47455", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "reproducibility", "evaluation_metrics", "documentation"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3a5cb8c0c7b6152679f37fa7c6f5c8a26e968001"}
{"id": "train_47456", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["governance", "reproducibility", "ci_integration", "auditability"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ed4f8ae03e72439d302ece9a9da0252c7ec58a25"}
{"id": "train_47457", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "governance", "documentation", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "18f562ed9b7c88fde4d64c9efe1ed5c42decfd8b"}
{"id": "train_47458", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "governance", "documentation", "auditability"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e3212f7a2f987dbfd9968a44af416a599b02fd51"}
{"id": "train_47459", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "tooling", "reproducibility", "governance"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0ff16158788839e712049971203b062235ef4c89"}
{"id": "train_47460", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "tests_are_truth", "ci_integration"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "522670aa2ad7d44f512b66e2fbe120150f048d8f"}
{"id": "train_47461", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "auditability", "documentation", "ci_integration"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5b81cad7912ec54d5cf01837d1956fae251482ce"}
{"id": "train_47462", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "ci_integration", "security_gates", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e991eb3aa628189970f204a325985d27a828e47f"}
{"id": "train_47463", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "tooling", "security_gates", "auditability"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c3d884570ac6bce9e536b3f9347a3ff4b7a0596f"}
{"id": "train_47464", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "governance", "auditability", "security_gates"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9466ddf7c0b3e5e7fb0ecf536cbe5e6676f4f0c9"}
{"id": "train_47465", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "auditability", "ci_integration", "security_gates"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8e687fe1b9666bd648373bd377e49d75ed01a9f4"}
{"id": "train_47466", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "tooling", "auditability", "ci_integration"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "128fef916cd17486a4b0eeca9515b7097f697a33"}
{"id": "train_47467", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "governance", "reproducibility"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e37905028b64951e5f1fffc110755a26d96421e0"}
{"id": "train_47468", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "repo_scale_reasoning", "governance", "documentation"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "54031425e0e64b261b2d535a77ab6bd15b6a7473"}
{"id": "train_47469", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "tests_are_truth", "evaluation_metrics", "tooling"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c9894db08ff55327f4957157eb45d85e80823bdd"}
{"id": "train_47470", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "reproducibility", "ci_integration", "tooling"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f1cd3297e9de2e0a13bb0a2a4d93e14b5190e4ae"}
{"id": "train_47471", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "repo_scale_reasoning", "tooling"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "17694da898e7b52f91f791323a8275559963a0b2"}
{"id": "train_47472", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "documentation", "security_gates", "ci_integration"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d8cad3d548047a925e24f4fa651fc4e2178faed9"}
{"id": "train_47473", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "documentation", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b353c2913e5d701928871a8bd4cb8a1a6c4ef16f"}
{"id": "train_47474", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "governance", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9a9e156bc5c95c12ea263b4c50c91744e551e2e6"}
{"id": "train_47475", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "auditability", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ceaea7dca75451790b9d650109ed26ddce17c153"}
{"id": "train_47476", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "auditability", "tooling", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fb4eb35266139c98bfaf952074468d395df199e8"}
{"id": "train_47477", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "documentation", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "059b9cb867b8caaea0594431ab041c70a091171f"}
{"id": "train_47478", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b2740e998ae6d8483d6deb036de9a7be35df20d0"}
{"id": "train_47479", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "governance", "tests_are_truth"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a438e4c2ac061963d3f1f9753dafc291d6176c1d"}
{"id": "train_47480", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "reproducibility", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "baa166d8ac42e06fe68db04154b9e49ee144860b"}
{"id": "train_47481", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "reproducibility", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8caceb915181228be977f49aa2bb6162cff171cd"}
{"id": "train_47482", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "auditability", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "14ebfd50b60d314517399ec8fd9b48fee719a235"}
{"id": "train_47483", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "governance", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "85cd073bcd306bd825d2f542253d056867a3f37e"}
{"id": "train_47484", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "cost_latency_tradeoffs", "ci_integration", "security_gates"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5bd73341c693fd9fb6b0945c86544dfc02324a1e"}
{"id": "train_47485", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "documentation", "ci_integration", "governance"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b4e5510362efcb7b042b42d1db4818e0745c345a"}
{"id": "train_47486", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "tooling", "documentation"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "28335e8910c5f5a7b8b1e1785fd736f639ce0e22"}
{"id": "train_47487", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "ci_integration", "governance", "reproducibility"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "082296cc5bd92efebd189dcce3be5ea9e0e52335"}
{"id": "train_47488", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "governance", "ci_integration", "tooling"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f0e86444332bf0d5858d609ab5b5ad0f3cf355f9"}
{"id": "train_47489", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "documentation", "reproducibility", "security_gates"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "153baa39f7eda2212582fbbb51706d65736d04f6"}
{"id": "train_47490", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "auditability", "reproducibility", "governance"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f394b9bfb36f5ec399857b242ab10bf5ae689c62"}
{"id": "train_47491", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "auditability", "security_gates"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a40cb459224ae5640510babf1701ab8fda9c7841"}
{"id": "train_47492", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "evaluation_metrics", "documentation", "ci_integration"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7120544bdc3c9ed1f2309fe943670d2875c82d91"}
{"id": "train_47493", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "security_gates", "evaluation_metrics", "auditability"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2084a496545a43f20c292864964e74e69f49af0f"}
{"id": "train_47494", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "auditability", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "51a11675b4682715d495495f667c06d81c945e1c"}
{"id": "train_47495", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "documentation", "reproducibility", "evaluation_metrics"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ccc1edc58bf4a70b3edf8e91f1dd493245d600ff"}
{"id": "train_47496", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "governance", "auditability", "tooling"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "460736a27f3ae3ab128a91c5bddf0ba4b1f20dea"}
{"id": "train_47497", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "tooling", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3d63f3693054699a2109df4bae64a68d6921126d"}
{"id": "train_47498", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "evaluation_metrics", "documentation"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4cbfe914385610b18328fc85b67ed98f5b9f2349"}
{"id": "train_47499", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "security_gates", "repo_scale_reasoning", "governance"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f593188eaf1f3ad4120fe7e8b80387e60fb935dc"}
{"id": "train_47500", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "reproducibility", "governance", "repo_scale_reasoning"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a350f78821fb775d4994038d511af086f26dabc6"}
{"id": "train_47501", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "reproducibility", "security_gates", "ci_integration"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "39c93383f5dd8807ddfb4f8f59d019eb2fac15e4"}
{"id": "train_47502", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "security_gates", "documentation", "ci_integration"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "98839e2634cada7c50404f3ef0206fdb128acc01"}
{"id": "train_47503", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "governance", "reproducibility"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1652adfc42be79ae3b26ae58ddc276765388f5e0"}
{"id": "train_47504", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "ci_integration", "security_gates"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d772a6290b841a6b4b4882b6058b5c5fb048a425"}
{"id": "train_47505", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "auditability", "security_gates", "reproducibility"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6f1b0822b3ab47d89c5a64be3f28479c1ecd923c"}
{"id": "train_47506", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "auditability", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9cd1e5d14987df743152de1f33787452a481d826"}
{"id": "train_47507", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "cost_latency_tradeoffs", "security_gates", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b7738fcc0eaed121fece69da87b6f262b5c02e70"}
{"id": "train_47508", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "tooling", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4e2c00c4843ceb175b78b3b6b8f7fd3d21a2a6e9"}
{"id": "train_47509", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "tooling", "auditability", "reproducibility"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a8c3ee8e052abe1635f0ae59db99c6073ff08abf"}
{"id": "train_47510", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "governance", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "780ba11b521df7e73a97ead9393aeee1444e5e1c"}
{"id": "train_47511", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "security_gates", "tooling", "tests_are_truth"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "563e53409d48836a88dd16b7467254b4e27725f2"}
{"id": "train_47512", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "ci_integration", "tooling", "documentation"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6e716811a15bb8ea27ffb050c60db85ddffba48b"}
{"id": "train_47513", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "cost_latency_tradeoffs", "documentation", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a82c53b10dc1cba33804c7ee146cf8e4b03c335d"}
{"id": "train_47514", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "ci_integration", "evaluation_metrics", "reproducibility"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "727eac7de664780600e9bfd8b675ea346231d53e"}
{"id": "train_47515", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "reproducibility", "ci_integration"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e34003a28d0f2eab6d7875e1e142b04849bbfee6"}
{"id": "train_47516", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "ci_integration", "tests_are_truth", "tooling"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "92d925411de9e441c577cdcc729796d9236dbfbf"}
{"id": "train_47517", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "auditability", "reproducibility", "ci_integration"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b9e1b652ca48644fa41a440cfc7aa16573f9c3a9"}
{"id": "train_47518", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "evaluation_metrics", "tooling", "governance"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "825e4c254fcc635367b6d2fcceb85337ed90c5b8"}
{"id": "train_47519", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "auditability", "documentation", "tooling"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4be3712f9995664f320d2699227e29d9fe2bd99b"}
{"id": "train_47520", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "tests_are_truth", "ci_integration", "evaluation_metrics"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d376ef0fb8b50840debf0af6167c1c5b06b8954e"}
{"id": "train_47521", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "governance", "tests_are_truth", "security_gates"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2878dd91d5796bc9b1a17ef72e6b5e963ee21754"}
{"id": "train_47522", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "reproducibility", "governance", "security_gates"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c14222e2d95ab5a53e0920ca5d8372680e4dfaf3"}
{"id": "train_47523", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "tests_are_truth", "auditability", "evaluation_metrics"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5ffa84ee14c0446342a78125b380bec3fc92dd6d"}
{"id": "train_47524", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "reproducibility", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "43db73511978e03a66fd24f3a944050c8d936873"}
{"id": "train_47525", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "auditability", "tooling"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f4c14cc0fc365959d7a90afa572002c492af50eb"}
{"id": "train_47526", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["governance", "tooling", "reproducibility", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dfad2fe9d224cc68a0c35b5bc978980166106484"}
{"id": "train_47527", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "security_gates", "ci_integration", "auditability"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b940aa023275f06942fa491a96d5e90e9fa99cab"}
{"id": "train_47528", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "security_gates", "ci_integration", "reproducibility"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0cd4e8b3e35fce4fbe4980d6c698f1a887bafad9"}
{"id": "train_47529", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "ci_integration", "tests_are_truth", "documentation"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c46bf7619015e4ffa107dcf8c65bf84568396743"}
{"id": "train_47530", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "documentation", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5764d26b06a04e627c200a338edfbe25edad3374"}
{"id": "train_47531", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "documentation", "auditability", "security_gates"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2dcc7dd4e77e13279ae51e82201a3d52d7f01cf9"}
{"id": "train_47532", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "auditability", "security_gates", "tooling"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "25aa2250d710a2007491305b9539e6ae31e4a4d6"}
{"id": "train_47533", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "security_gates", "governance", "documentation"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f2bd519b60c0afa8e1de002ba5fd748a622e1765"}
{"id": "train_47534", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "tests_are_truth", "ci_integration", "auditability"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d6c024ed744735eb47e0a9aa8d821affe2a1c05a"}
{"id": "train_47535", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "documentation", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6a5330be73c8ad7c9effa3647b1e6da76d1e66b5"}
{"id": "train_47536", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "evaluation_metrics", "reproducibility", "auditability"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ffb5af248f6d71fdc182b1995cd227b257e4599b"}
{"id": "train_47537", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "tests_are_truth", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "166d2b9329d42404129d039839b7d14dc12febf3"}
{"id": "train_47538", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "documentation", "tooling", "security_gates"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a53150fc9fea421c3bef73e76f9adf88d5944ca1"}
{"id": "train_47539", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "cost_latency_tradeoffs", "security_gates", "ci_integration"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "907baf22d69990a88f6bb82094d6f017cb71e8f6"}
{"id": "train_47540", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "tooling", "reproducibility", "documentation"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "264aa07c9b9b6e5ee6471c3c5bf019a98ddcb0ee"}
{"id": "train_47541", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "ci_integration", "tooling", "auditability"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c7c722963563fab13abe34f5b06844451af55cb0"}
{"id": "train_47542", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "governance", "ci_integration", "reproducibility"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "81808e553142096e1e5584dcfe77ba318ddb414b"}
{"id": "train_47543", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "documentation", "auditability"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a7c88598c70878d989f4f474e3ba7edf0eb59754"}
{"id": "train_47544", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "security_gates", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2be88a2f9a411c24e21c9b8e392699eabfef58d3"}
{"id": "train_47545", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "tooling", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "42b43ebfb3506368667e1f0abe23c7551a258817"}
{"id": "train_47546", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "documentation", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "38dd9970493480e9438f95a0c842a2ec48ee6678"}
{"id": "train_47547", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "documentation", "tests_are_truth", "governance"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cde3c7d6acd3a71be6d0269d2b46f754218538b3"}
{"id": "train_47548", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a9b6a5f9c7d4c019a68688162777a2521177a970"}
{"id": "train_47549", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "ci_integration", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ec269ecf28a893834f7b228ccd2b3353521be186"}
{"id": "train_47550", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "security_gates", "tooling", "tests_are_truth"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "de8b87756832be47d89ca31867bfe4bc7130b44f"}
{"id": "train_47551", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "evaluation_metrics", "ci_integration", "tests_are_truth"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d131bb3522c1ff551a7bf3ec74d60337b8092183"}
{"id": "train_47552", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "documentation", "tests_are_truth", "reproducibility"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3707eaf7d2814f3d8599dac4512f88c8ea00c4ec"}
{"id": "train_47553", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "evaluation_metrics", "repo_scale_reasoning", "tooling"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3e84ffe4360b8c72447035fb4bdf77957e4cd6b3"}
{"id": "train_47554", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "ci_integration", "reproducibility", "auditability"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e5e454507274b4b95a9168a33914b61c39ade27d"}
{"id": "train_47555", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "evaluation_metrics", "repo_scale_reasoning", "security_gates"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "665099c93d14093c7e31f869b11f8ed6583430d3"}
{"id": "train_47556", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "documentation", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "96e71c02541f187371acada127ef3d91176addc2"}
{"id": "train_47557", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "auditability", "tooling", "reproducibility"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b5b262ea8a1978dc1c179d00dd9a05ed292c9c2c"}
{"id": "train_47558", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "security_gates", "repo_scale_reasoning", "documentation"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "81585c228021d39a6236ac567581dab83f9f8a8d"}
{"id": "train_47559", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "evaluation_metrics", "reproducibility", "governance"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fae6758a4cea977c8d3dc2de4146ab18b0a0b444"}
{"id": "train_47560", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "tooling", "repo_scale_reasoning", "documentation"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "47f97d79a6eec426ec99258892dba3f094733a4b"}
{"id": "train_47561", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "reproducibility", "documentation"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9397d095e22a584736b61ba239696a05ec384d54"}
{"id": "train_47562", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "security_gates", "governance", "reproducibility"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "049594df47c84809364b96b777dd2d9ad7de460d"}
{"id": "train_47563", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "tooling", "tests_are_truth", "reproducibility"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7165ffdf2d0915b2ede3cdc01e12a7e90d9d39d1"}
{"id": "train_47564", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["governance", "tooling", "security_gates", "tests_are_truth"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5069932daa1bc90f0d208b27af2e95e039cea86c"}
{"id": "train_47565", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "ci_integration", "governance"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1bd4e3762ac92d6b109da2cf232f9cbcf71bbaac"}
{"id": "train_47566", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "security_gates", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e2bcf02d2c096b6337567adc48b26059269c4ceb"}
{"id": "train_47567", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "repo_scale_reasoning", "reproducibility", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "95df3e09de11387d6fc7a3c13041904b75459a79"}
{"id": "train_47568", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "auditability", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "57d151272b74c75c58bb6503388c9f8a16488221"}
{"id": "train_47569", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "cost_latency_tradeoffs", "governance", "tests_are_truth"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d722b93bbfe7c2d8cb5914a96508b5cbd11c1a47"}
{"id": "train_47570", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "governance", "reproducibility", "documentation"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ba2c046fb5bea3a38a2cc19193400bd4f176ce4b"}
{"id": "train_47571", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "tooling", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c690acbf53690aa9f2a2bae9251b2886c5f762b3"}
{"id": "train_47572", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "documentation", "security_gates", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "97cdb7c198a1697f91664b8bc04827ff585d9ea6"}
{"id": "train_47573", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "ci_integration", "governance", "evaluation_metrics"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c5ecebac2cab0654cf9eb6f492d4949f654ec599"}
{"id": "train_47574", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "documentation", "governance", "auditability"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8629bc1e099d279efafa5d1562f25352bc1e40c8"}
{"id": "train_47575", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "tests_are_truth", "security_gates", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d4ee6163e373b8c9d857c0eba9eb49912bd83f6c"}
{"id": "train_47576", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "reproducibility", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "096cfe53af6c5aada085750e1471826cb793151b"}
{"id": "train_47577", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "reproducibility", "documentation"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "70d10286eacf298fc2cc18f680aef7fb92c7bb3b"}
{"id": "train_47578", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "security_gates", "auditability", "documentation"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b23deb30c46616b98ad199c00744c5096008ce39"}
{"id": "train_47579", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "repo_scale_reasoning", "auditability", "governance"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ccb14d32989bf8e7e49d0e31219366f52e6e9730"}
{"id": "train_47580", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "tooling", "tests_are_truth"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4cbfe77fd04e38108d7d445c4a53d0554f6d8182"}
{"id": "train_47581", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "governance", "repo_scale_reasoning", "tooling"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eb10dd3d892dcfae7c2505d6f8f2e7011cae3016"}
{"id": "train_47582", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "security_gates", "reproducibility", "documentation"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0698fa67edd5751fbf83251bd1b07aae790c155a"}
{"id": "train_47583", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "documentation", "tests_are_truth", "reproducibility"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6f4075effe229aa27c83bbd0f74d6f0e0d904a28"}
{"id": "train_47584", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "governance", "documentation", "evaluation_metrics"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "19009fdb6f1b59f34c5eda5d4931a02e54b6499e"}
{"id": "train_47585", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "evaluation_metrics", "reproducibility"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "16f6e237461e76149fb67a9214e2b017edb309e3"}
{"id": "train_47586", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "governance", "tooling", "security_gates"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "95e3a60a594cc77f0e9a51cf331ea965ac3cf12c"}
{"id": "train_47587", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "security_gates", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "56598f526666cb03f3148b74a8e20b353b66d3e6"}
{"id": "train_47588", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "governance", "tests_are_truth"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "39ceb873ddcd250b3df54f7af3be934bff6cf816"}
{"id": "train_47589", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "security_gates", "ci_integration", "auditability"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7330ac8b06ca156ea6fe32ea66ca71c5c31462c2"}
{"id": "train_47590", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "repo_scale_reasoning", "evaluation_metrics", "security_gates"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3faebf18c7c4223ca96a8dc65e28f7cbcf679796"}
{"id": "train_47591", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "governance", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e9492808c8825ff19f183cfffa2af785b37185ec"}
{"id": "train_47592", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "evaluation_metrics", "tests_are_truth", "governance"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "628156ae0be9347d2e8d664e126abd3870767a7f"}
{"id": "train_47593", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "governance", "tooling", "tests_are_truth"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f3bd8ec65917d76745d3931754fb6f48e341a966"}
{"id": "train_47594", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "auditability", "ci_integration"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1753ad5d5602be652ee03386292828a85c7b19cd"}
{"id": "train_47595", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "governance", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2e4037738061c37389b316435cc27adc38918a59"}
{"id": "train_47596", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "reproducibility", "governance", "auditability"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3fa508a9aaf69ff95b1f42e7e0b4e6d604d3d4c3"}
{"id": "train_47597", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "security_gates", "tests_are_truth", "tooling"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "08336c5bdee75f83c243b3bff4ce21d7e2e4e9c2"}
{"id": "train_47598", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["governance", "tests_are_truth", "reproducibility", "auditability"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5f99d9680da1414912bfd8618dbf4ebd3c671c23"}
{"id": "train_47599", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "tests_are_truth", "reproducibility", "evaluation_metrics"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "acc4f570e322ef62f6d0cff06232f56f081f5461"}
{"id": "train_47600", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "evaluation_metrics", "governance", "reproducibility"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "34ea5ab3789941bdbd5a16636018459acb529f1c"}
{"id": "train_47601", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "auditability", "tooling", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "703feec917b7fb21a47f57b40d0e8fbee9cfc8fa"}
{"id": "train_47602", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "tooling", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5fd5f26bb205f3257f16f534d3f765b1d8fc072f"}
{"id": "train_47603", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "documentation", "repo_scale_reasoning", "tooling"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d58c0a4006eb35d911fee88957a700321e9b1944"}
{"id": "train_47604", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "evaluation_metrics", "documentation", "auditability"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "001c6f647411889a75d75ea1c41d2019b08ffae4"}
{"id": "train_47605", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "tests_are_truth", "repo_scale_reasoning", "documentation"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "80bd424438f1e9f1362113d4ebb75687d55a9a54"}
{"id": "train_47606", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "security_gates", "tooling"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6accc5c864f1377519675abea3537f7eb645ffd5"}
{"id": "train_47607", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "auditability", "security_gates"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8a209398280ddc15f8cd997f45911fbd1eee3d03"}
{"id": "train_47608", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "security_gates", "governance"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4173460e45a444d22d4de72a547f586097d82a35"}
{"id": "train_47609", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "security_gates", "ci_integration", "evaluation_metrics"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b4137cfba47f222c44ff04a077073b8bce8610bd"}
{"id": "train_47610", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "governance", "ci_integration", "documentation"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6c7bcafde229e02a9a92f40bbe34518df72f202f"}
{"id": "train_47611", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "governance", "auditability"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "554a165ebd8fd2e44f095a041e8e77beea5df4ae"}
{"id": "train_47612", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "auditability", "documentation", "reproducibility"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2e51e76a696fed8c557f91305aded37cda86ea90"}
{"id": "train_47613", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "auditability", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "de1f0da187bc60c90d7d4322365effeb8744766c"}
{"id": "train_47614", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "ci_integration", "security_gates"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1ebffcc1c0130d61863a85c22c6939e5963874b3"}
{"id": "train_47615", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "documentation", "ci_integration", "governance"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "45497913664d77503e321794f003a260b6524a76"}
{"id": "train_47616", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "security_gates", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "330d497bb66e23d821cc31f417072cb95ab7c33f"}
{"id": "train_47617", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "auditability", "governance", "tests_are_truth"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5c66fa4110e451e61116258de3b960a865e0cf3c"}
{"id": "train_47618", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "governance", "reproducibility", "security_gates"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "33f80798321daf1dafa9d753b7e7c422866b5a7a"}
{"id": "train_47619", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "ci_integration", "evaluation_metrics", "security_gates"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "112f6f914d6309814c6e25adc2dfa80608a9f908"}
{"id": "train_47620", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "auditability", "reproducibility"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f04116d361708962a39ae4885c62243b98e8b9e2"}
{"id": "train_47621", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "ci_integration", "evaluation_metrics", "documentation"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "743c3b3ae1be84e14a930963c964e85a2e1ffcc3"}
{"id": "train_47622", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "reproducibility", "governance", "evaluation_metrics"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7c6856c5871bc8b9f46df5cf33fa1c26a142fcc0"}
{"id": "train_47623", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "ci_integration", "documentation", "governance"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a6ed3ddca2981feb0a13bfe7311ad3ef1ab40e8d"}
{"id": "train_47624", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "reproducibility", "security_gates", "ci_integration"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fcddeae68574cc57640c4e3c3f772cda7103a735"}
{"id": "train_47625", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "documentation", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "293534e855c285f83d0dd6887b8ea828b7079bed"}
{"id": "train_47626", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "ci_integration", "evaluation_metrics", "auditability"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "24ec06c71883dc99b78ee58f8e72880d3908963e"}
{"id": "train_47627", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "tests_are_truth", "security_gates"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "acef2d4f6a522a6353fe5fe7c7f5523f80796541"}
{"id": "train_47628", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "documentation", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0b3c7800daacf6f26c9e87f6176a352e47e8c0ae"}
{"id": "train_47629", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "auditability", "reproducibility", "tooling"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b658a01f35ab93c6756adda5bcfb5ae2bbf1f34f"}
{"id": "train_47630", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "ci_integration", "auditability", "reproducibility"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5bdd028724f41f397e33a527545d6bd04e89c619"}
{"id": "train_47631", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "documentation", "evaluation_metrics", "auditability"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dde973614d1f4a1e1ec65e368070e64e332e08ce"}
{"id": "train_47632", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "ci_integration", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "53b256c53d05e666131c3b93e1fb857a05136598"}
{"id": "train_47633", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "evaluation_metrics", "repo_scale_reasoning", "auditability"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f646e3d72cf1c8e891ee2c874c8bc13bb18fbb27"}
{"id": "train_47634", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "tooling", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "399be04e1c87abdf984fbd8a59fdfcad0149ea8f"}
{"id": "train_47635", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "documentation", "reproducibility", "tests_are_truth"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a0300692b8eb8e4cf702460ae991d761b190dca5"}
{"id": "train_47636", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "tests_are_truth", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "54b10f94ca99902ca1e3a4e3578090d97dd080e6"}
{"id": "train_47637", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "reproducibility", "auditability", "ci_integration"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dd1a692dfe8610ffd9885e5a616426249f4c33fb"}
{"id": "train_47638", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "repo_scale_reasoning", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "725164ec12fe12c798d02454767cd2f9a48df20d"}
{"id": "train_47639", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "tooling", "ci_integration", "auditability"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ce6e3e5c3362d0f7f492dfd5efc27cf42c963843"}
{"id": "train_47640", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "governance", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "14a56e6961408aab2f1c66e8a20b02ff8c3a0a9b"}
{"id": "train_47641", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "auditability", "ci_integration"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7ebeee507d1099954739d0f31b7861c768cbf232"}
{"id": "train_47642", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "documentation", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8ec460f8128461acde45bca1ff9bc81997c3b53d"}
{"id": "train_47643", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "ci_integration", "tests_are_truth", "reproducibility"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "473c70a7f2f0c744b2f86dda9ecc3483d5050a21"}
{"id": "train_47644", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "governance", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "79c710cf6c44f068d11025e43dab312be40cbe05"}
{"id": "train_47645", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "governance", "evaluation_metrics", "documentation"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9d73c766459d00f37bf000e6d84456d42330fd60"}
{"id": "train_47646", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "auditability", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "321d4cfcd9e0918650f77b9be4790a2cc744c493"}
{"id": "train_47647", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "documentation", "reproducibility", "security_gates"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "802f198054e470f648ac04bd9d8fc12d398d09bb"}
{"id": "train_47648", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "evaluation_metrics", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7ada061021d94e5b805fdb57cbed064e130b686d"}
{"id": "train_47649", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "ci_integration", "auditability", "tooling"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7bc9ae22620cf5a05ed97f8b39d3487b7d6ab34a"}
{"id": "train_47650", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "security_gates", "reproducibility", "tests_are_truth"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e761b8f64efbae9a7dddffa8bb831aa305626fc1"}
{"id": "train_47651", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "security_gates", "documentation", "ci_integration"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a7bd47632f6d63fe74e3271c1a905bc05705e0bf"}
{"id": "train_47652", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "reproducibility", "evaluation_metrics", "documentation"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3cce97c47af17e79656701c1d0636c846e0fbca6"}
{"id": "train_47653", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "governance", "documentation", "auditability"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "350c1d7e92ee380a5cace7ff37d760e152209423"}
{"id": "train_47654", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "cost_latency_tradeoffs", "reproducibility", "governance"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c3c4c2115b34bf3dab4dcf6fe9d2d2ca4812c2a9"}
{"id": "train_47655", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "security_gates", "repo_scale_reasoning", "documentation"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0e8e31749d19358df370009266c32ebbc55dcbc5"}
{"id": "train_47656", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "security_gates", "auditability"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bebd13c45819a05f2be28d6ac219a8711e100bfd"}
{"id": "train_47657", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "governance", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "580877fab905b42c6e8f46e56582ae593979b1f9"}
{"id": "train_47658", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "tooling", "security_gates", "tests_are_truth"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eff2337e41c2c2255cbcf300afa97d3b9d397f05"}
{"id": "train_47659", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "documentation", "tooling", "security_gates"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3c67a039c5dc626f7703afdb5dbbd1917e9379a6"}
{"id": "train_47660", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "tooling", "security_gates", "evaluation_metrics"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9d9a62b37c1c9f2ffa9d6a702ee7232ce74904d3"}
{"id": "train_47661", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "tooling", "security_gates", "tests_are_truth"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2ac837a1554111a9c6fd427c142fa90f41424f6b"}
{"id": "train_47662", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "cost_latency_tradeoffs", "security_gates", "evaluation_metrics"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2093c0b9d150f01ef11a898da3f13923f9105e1f"}
{"id": "train_47663", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "evaluation_metrics", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "87e39b05d6506fd6dbb920befe2bb668caeee720"}
{"id": "train_47664", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "repo_scale_reasoning", "ci_integration", "evaluation_metrics"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ae2186d227cf86cbbd88f90129b5dd6eff24377c"}
{"id": "train_47665", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "evaluation_metrics", "security_gates", "ci_integration"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f268a2fd4dc135af481794d30b9a00e0a5fc40bf"}
{"id": "train_47666", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "governance", "reproducibility", "tooling"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8c602db113e5ff6725ad62e3279150edc6615f99"}
{"id": "train_47667", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "repo_scale_reasoning", "governance", "tooling"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "67aa8431dcfd32f23559c4dfe1844d1982e214cf"}
{"id": "train_47668", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "evaluation_metrics", "repo_scale_reasoning", "tooling"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3543ea9e9b28cace873220a9d326c35e7d8711f2"}
{"id": "train_47669", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "repo_scale_reasoning", "auditability", "security_gates"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f585728dca7be63143ab851823ccc1ac6074eb56"}
{"id": "train_47670", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "documentation", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e650a6101bb8dfacea37eb37f08a27a85a31b6ac"}
{"id": "train_47671", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "security_gates", "governance", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0845d7f12152810a0096d7b025299e30ebf388e3"}
{"id": "train_47672", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "reproducibility", "evaluation_metrics", "ci_integration"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "992f567fa5d6d2b5a6870bed9256d6371dca7726"}
{"id": "train_47673", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "repo_scale_reasoning", "tooling", "governance"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "54cce7c54971f0734ae41062b8541afb78e7d4de"}
{"id": "train_47674", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "tooling", "documentation", "security_gates"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0930f012a46ea5df3989d73625104fbc114bf1c4"}
{"id": "train_47675", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "repo_scale_reasoning", "tooling", "reproducibility"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c3bd8e610cc98a9d7b85a4d01346af84afa3c483"}
{"id": "train_47676", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "tests_are_truth", "auditability", "reproducibility"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9f004a01f9a01e59effafa556f3ee8967243270d"}
{"id": "train_47677", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "repo_scale_reasoning", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "66891ab1cc4b4f469207a3159c5fbbbde46eed82"}
{"id": "train_47678", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "reproducibility", "evaluation_metrics", "ci_integration"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "325de52976c5b2345d9456b94259b97044cdaf72"}
{"id": "train_47679", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "ci_integration", "documentation"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "99fcc21861df122f123844f0eebe57aaea3d972f"}
{"id": "train_47680", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "governance", "auditability"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "84bfeb9180b707f48f23c4e974231b87800d69bd"}
{"id": "train_47681", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "repo_scale_reasoning", "tooling", "documentation"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "06dc65d1e1cacccc00f52533b9bb6937c53ffef9"}
{"id": "train_47682", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "ci_integration", "tests_are_truth"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a40a49cc03bb4aeba7185f6c9d791035aef1b164"}
{"id": "train_47683", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "auditability", "tests_are_truth", "tooling"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7e232a1a9b492ba40c4fd122907a214b64b3e7f8"}
{"id": "train_47684", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "security_gates", "auditability"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8cd37b4842a8fdc3d85152de7ccd6c041fb7e1fb"}
{"id": "train_47685", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "governance", "reproducibility", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "67a4748e8a39ea4796abbcbd8fd0e9f3aa247cc0"}
{"id": "train_47686", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "reproducibility", "auditability", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d31b606693ae7a0d12af5f473016587b169bf86f"}
{"id": "train_47687", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "tests_are_truth", "governance", "auditability"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b6617ab6d29f1670f0dff61efa82c8d3bf10d600"}
{"id": "train_47688", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "evaluation_metrics", "security_gates", "ci_integration"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a1415668da0b9ad519dc3ec03157e5e956c388fb"}
{"id": "train_47689", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "ci_integration", "evaluation_metrics", "security_gates"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b153f8425d90617d621708060086d02d70ed4160"}
{"id": "train_47690", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "security_gates", "reproducibility", "documentation"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "864c8afc44cc91ed9979c171a3d3b2d29a24dd75"}
{"id": "train_47691", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "ci_integration", "documentation", "evaluation_metrics"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ab99bf5ba11eececdc01a08d59e29d28cb3c6c8b"}
{"id": "train_47692", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "governance", "auditability"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "71d4d1c78e012139534f211bcbd50c9f0357b6d0"}
{"id": "train_47693", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "evaluation_metrics", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b4e98a1555660e03fe2c8aca2dba5809b13e90b4"}
{"id": "train_47694", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "repo_scale_reasoning", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d679c205be01274bc44238e04e8fc9a327c53c6b"}
{"id": "train_47695", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "documentation", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4ac8a8df5aa8659606bcd94700d25400c1a290f2"}
{"id": "train_47696", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "auditability", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6d000dcbc0b195611c1aa5fb728bc35540ca534e"}
{"id": "train_47697", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "security_gates", "tests_are_truth", "governance"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "75fccde20fc7b45756c0aebd6b85240d9279ed17"}
{"id": "train_47698", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "governance", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c4b9b8b499766f44f40429df69902de452555b6a"}
{"id": "train_47699", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "tooling", "governance"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "004b49ccf8eccf02c41b247910da1b090e47f9ca"}
{"id": "train_47700", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "ci_integration", "auditability", "tests_are_truth"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f4011e6efaba616304991d840a2559bda0a35512"}
{"id": "train_47701", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "auditability", "ci_integration"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aa80581c6a84ad2da7f2256cb7cd99990822d2c5"}
{"id": "train_47702", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "tests_are_truth", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "563e53409d48836a88dd16b7467254b4e27725f2"}
{"id": "train_47703", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "auditability", "reproducibility", "documentation"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eb32dc2924139868ae3b62b26b5144b7a11c56f6"}
{"id": "train_47704", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "security_gates", "auditability", "repo_scale_reasoning"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "41b92495c66e5021d19fa6fb15f9a226017bec42"}
{"id": "train_47705", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "security_gates", "tests_are_truth", "ci_integration"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fcff38b70336ddf0022319da88c827117153aa7d"}
{"id": "train_47706", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "security_gates", "tests_are_truth", "documentation"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "53fe9a3d4acfc489ba4244ff76fa3d33dccab937"}
{"id": "train_47707", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "ci_integration", "reproducibility", "governance"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d64ded8315be8d18059a8439daf049670c61fee0"}
{"id": "train_47708", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "evaluation_metrics", "reproducibility", "auditability"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a82c53b10dc1cba33804c7ee146cf8e4b03c335d"}
{"id": "train_47709", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "ci_integration", "tooling", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cb65172e824691ebaf82f7696b5638687659cda1"}
{"id": "train_47710", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "governance", "repo_scale_reasoning", "documentation"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "806f061d05cff75d46605638ab923c2a33f5a0bc"}
{"id": "train_47711", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "tooling", "documentation", "auditability"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1360d95f676cb2003a6afec1f9a8be5ceec8c5c8"}
{"id": "train_47712", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "evaluation_metrics", "ci_integration", "governance"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "164043cd5589f20c34404ee4495b5297d8072087"}
{"id": "train_47713", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "tooling", "ci_integration", "governance"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5976639c522c4bcb35f7f7e15bf16b06ae114bdd"}
{"id": "train_47714", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "governance", "ci_integration"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "94b13a16538b57baf55c0634176a9013dc5edbfe"}
{"id": "train_47715", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "governance", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b57c89c3990811fbce53348e889e7c736556deb5"}
{"id": "train_47716", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "ci_integration", "reproducibility", "governance"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8c7720c80cd1fee8b94176decba15bebda5d19d4"}
{"id": "train_47717", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "tooling", "governance", "auditability"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ba735827a3359d0ef083253102a834b0efaa0e47"}
{"id": "train_47718", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "governance", "tests_are_truth", "reproducibility"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5004ef889aee62ec0b538c3114589374b068135d"}
{"id": "train_47719", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "ci_integration", "documentation", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0606ab52886c8b53f0d0b63bdd80ebb25a810ec4"}
{"id": "train_47720", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "tooling", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "92a6b00788a1988bd778236ff39a45ad5ec52557"}
{"id": "train_47721", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "ci_integration", "auditability", "security_gates"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e02079da83184b185d58edb85424f9d4c58611c6"}
{"id": "train_47722", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "reproducibility", "auditability", "tests_are_truth"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f346aaf08f021fb592e01985fae800c81cf4fbf9"}
{"id": "train_47723", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "auditability", "security_gates", "ci_integration"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0b7e9f26a3e2d5fc4044523218f92441c48144f1"}
{"id": "train_47724", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "tooling", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e0b946571d7a368e38b02ec93d21af441a47959e"}
{"id": "train_47725", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "auditability", "security_gates", "reproducibility"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a889278fe733b9694cc40d6b0cd7944f85edfb61"}
{"id": "train_47726", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "governance", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "94bd31124501164e37273230d8a45851f6b60c46"}
{"id": "train_47727", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "auditability", "repo_scale_reasoning", "security_gates"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eac40d489f19f4495a56804352725b5201626633"}
{"id": "train_47728", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "auditability", "documentation", "tooling"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b7d4fe8466f415ad6fc660713c064e1a6aae3953"}
{"id": "train_47729", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "ci_integration", "documentation", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cf0bd9e3f8e6e3455a061061cc16f0b3ad95e986"}
{"id": "train_47730", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "reproducibility", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "777cec47d8aecdba531b235e17b25f0580d792cd"}
{"id": "train_47731", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "evaluation_metrics", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fdba311d7343e5d7701e9de961a3297908ea119d"}
{"id": "train_47732", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "ci_integration", "governance"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cac780c28178d9fada05067fc20f6b4259f4849f"}
{"id": "train_47733", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "repo_scale_reasoning", "auditability"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "98276aa26d5f05384e52c5eccda949bd67a52afe"}
{"id": "train_47734", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "cost_latency_tradeoffs", "repo_scale_reasoning", "auditability"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1d3bed88d17fc2d30ca2398ad7c23ce4d0ac6a80"}
{"id": "train_47735", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "ci_integration", "tooling", "documentation"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "67970d5a866fdefbf7c54573204c413fae65a998"}
{"id": "train_47736", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "tooling", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9d008e18e21e62705d8398da94f47dad02bd5705"}
{"id": "train_47737", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "governance", "ci_integration", "tooling"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3127dc5dae47b54b4e293830894743c266ad75fd"}
{"id": "train_47738", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "documentation", "governance", "tooling"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e7dbf6a4c4e4dbb7b4272cf72c7eeca84c72c29a"}
{"id": "train_47739", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "governance", "tooling", "documentation"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "89ef97748a4eac217bd3d775612edabc4cc04c05"}
{"id": "train_47740", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "tooling", "ci_integration", "tests_are_truth"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cff8ead94eec108b63238a06716d6c73722cbae0"}
{"id": "train_47741", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "documentation", "auditability", "ci_integration"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "162aeb95ed0c2b4b8f4639ab14644742e1064e4b"}
{"id": "train_47742", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "reproducibility", "auditability"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2340c3e78f53bfb37c8293540c6290e67d78148a"}
{"id": "train_47743", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "governance", "tooling", "ci_integration"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9d6b281c088180d455320e54b2290647f08a4157"}
{"id": "train_47744", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "documentation", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "489b4210a05cf21177fb343ba05b038c1aabec7c"}
{"id": "train_47745", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "auditability", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "74a386662ed22879cd00aec239650f8ea2aa8887"}
{"id": "train_47746", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "governance", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3613720c041a2a20ea97da6887c030ee7f5716ce"}
{"id": "train_47747", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "documentation", "tooling", "auditability"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "88cd211d7cefa9322cf3727027eb1a13d3f8ef13"}
{"id": "train_47748", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "governance", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "79a634cb297b4ce46300606ab5da34baf434cd35"}
{"id": "train_47749", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "evaluation_metrics", "reproducibility", "tooling"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e70d1e540ba2b6b449e70087020d28188f75e6fd"}
{"id": "train_47750", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "repo_scale_reasoning", "reproducibility", "ci_integration"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "22ddb0883242706233a4e9364738ad8376f5f1b2"}
{"id": "train_47751", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7ae359be7b8fa6c9ad5f25e9d18a1e3fd308339d"}
{"id": "train_47752", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "evaluation_metrics", "documentation", "ci_integration"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6150597f06bc7737bc8eb183f2238d0f3ae59b85"}
{"id": "train_47753", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "tooling", "repo_scale_reasoning", "auditability"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "026dfdc4a87c0f6eaee6a570e955194d970cf9d0"}
{"id": "train_47754", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "ci_integration", "repo_scale_reasoning", "auditability"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bb421a47ac1a579ad108e1250014f6c18199b618"}
{"id": "train_47755", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "security_gates", "governance", "tests_are_truth"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f2d101402ebfe2ce5bca718a0c88343a50a477b4"}
{"id": "train_47756", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "auditability", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "88cf237660f1d91d3a420d4445a9585d894fc50b"}
{"id": "train_47757", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "auditability", "governance"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0497b58493edb67906620bce10cc7eb24d41a562"}
{"id": "train_47758", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "auditability", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "850d831295998a7d918821d90bf56e2398720f00"}
{"id": "train_47759", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "evaluation_metrics", "repo_scale_reasoning", "documentation"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7afd2007fda6c77ba634218eb304a1c087f02d5a"}
{"id": "train_47760", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "security_gates", "tests_are_truth", "auditability"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6cea28f2029d0ffe0406682dc834b4d2f2f0cc6c"}
{"id": "train_47761", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "governance", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "de4c658580592d0b323ce286762fcbca93f4a494"}
{"id": "train_47762", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "ci_integration", "governance", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "85cd073bcd306bd825d2f542253d056867a3f37e"}
{"id": "train_47763", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "documentation", "evaluation_metrics", "tooling"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b4ff8b8a97f8fcd08529bfd5474f810b7b33d3f3"}
{"id": "train_47764", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "evaluation_metrics", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e0cbc00ea784253fa2892fbfa98674092f6226ff"}
{"id": "train_47765", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "tooling", "evaluation_metrics", "ci_integration"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1df8fecd705631b03dbefb5c1d400f23862b9bd2"}
{"id": "train_47766", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "auditability", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0775307aba77d57e4ba1da647e9e4ecf7faa07ec"}
{"id": "train_47767", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "auditability", "tooling", "reproducibility"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "85072c10e3569d40638489db59aedb0577100514"}
{"id": "train_47768", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "governance", "documentation", "security_gates"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3de822988aef77719f3d8ef9696707e5c954b158"}
{"id": "train_47769", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "ci_integration", "evaluation_metrics", "governance"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "31bb34a20aae8a557c9cb451f4d20abe55c3ec6d"}
{"id": "train_47770", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "documentation", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "35017f1bd2e88090bf2c84531755cae0c53e768b"}
{"id": "train_47771", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "evaluation_metrics", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5148af9f0bfe32d906baedb2d8693d38910fd4c0"}
{"id": "train_47772", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "documentation", "auditability", "tooling"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "80e2b4293e1d9bf99749f621dadee81a818ca231"}
{"id": "train_47773", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "repo_scale_reasoning", "reproducibility", "auditability"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ccc6b7592c47b722035bf0ccde2d0bf1e45ca5f5"}
{"id": "train_47774", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "governance", "tooling", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "20c328732f8964a0be1997dfee2fae9d8f3327ea"}
{"id": "train_47775", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "tooling", "auditability", "tests_are_truth"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "51b9b2afb1794e548be54b3920df58bc4b9787d5"}
{"id": "train_47776", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "governance", "repo_scale_reasoning", "documentation"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "01c681b1a5de98d750e2e4c82232df6113595ef2"}
{"id": "train_47777", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "ci_integration", "documentation", "reproducibility"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "51a9e02896e0e15ddf34624529576719d9d32d16"}
{"id": "train_47778", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["governance", "tests_are_truth", "documentation", "tooling"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4ed0f2c3c68ef84f2a6df5044dc6f2683be482f4"}
{"id": "train_47779", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "security_gates", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ffcb5f25765651769d5e72cbad55b1bce700d157"}
{"id": "train_47780", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "ci_integration", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "27d9dc790ea345b388bb1c40896fbda64be00d6e"}
{"id": "train_47781", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "auditability", "security_gates", "documentation"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "004208a7bdfe34ecc35bc8510d583daeb5fba804"}
{"id": "train_47782", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "tooling", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "17923cd4b616d74550014c0aeca8197e9f8e830c"}
{"id": "train_47783", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "governance", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "798aedff363a2bf6a6821c1b1d7c0714d6d5d3ba"}
{"id": "train_47784", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "auditability", "ci_integration"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fbc00b8699de181f76d978027ebd8ff1374d1f7e"}
{"id": "train_47785", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "cost_latency_tradeoffs", "security_gates", "auditability"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c19642c90d295fdffc7453b5eccc8e7a6cd3840c"}
{"id": "train_47786", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "tooling", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bdf5c04fc3c5c936fc25fd3054cb33e981626ee4"}
{"id": "train_47787", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "auditability", "tooling", "tests_are_truth"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b9793adc0a22813f132c9862d38ac6f69b8678aa"}
{"id": "train_47788", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "reproducibility", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5035575bb8fac13d795235622fda8ae6c003a4ea"}
{"id": "train_47789", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "reproducibility", "tests_are_truth", "security_gates"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fa2927fb317c67ffa7f0ae93ba17d654751202c0"}
{"id": "train_47790", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "auditability", "security_gates", "tests_are_truth"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "78fbe113496c8e83525d15d397303023711797fa"}
{"id": "train_47791", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "ci_integration", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cea35f669fbff98089b8487a8767e8632816acb3"}
{"id": "train_47792", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "repo_scale_reasoning", "reproducibility", "evaluation_metrics"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ea55aab5ca7e3a0a30a47ff8b4c9e20a8c145be6"}
{"id": "train_47793", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "repo_scale_reasoning", "reproducibility", "auditability"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "df5d80a9f0080230fd64951757c359a94244e591"}
{"id": "train_47794", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "ci_integration", "tests_are_truth", "security_gates"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "916be52fde9073e1b0e772ae49c4760fbf261c9f"}
{"id": "train_47795", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "security_gates", "repo_scale_reasoning", "tooling"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e56c8e50f1d02b310d800fe6f752eb8a22592142"}
{"id": "train_47796", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "repo_scale_reasoning", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "094fab2e0d8e8b0e2f0fe815ae1e0e22bf9a8034"}
{"id": "train_47797", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "auditability", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8b4fc65d69071d01247c50d3b41733a3cf144bd3"}
{"id": "train_47798", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "security_gates", "ci_integration", "governance"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "46c3539b58872c080acc8d3554f22e13173ad01f"}
{"id": "train_47799", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "tests_are_truth", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "df9b9fc841dfc993a9be78bfd3900fe15e33c84d"}
{"id": "train_47800", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "repo_scale_reasoning", "ci_integration", "documentation"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e9b59f99a0ebaae597b373709bc7c0f03a1cc3f0"}
{"id": "train_47801", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "ci_integration", "security_gates", "governance"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "527fc3df6954f080a1256fd19a8f3cc63b6482d5"}
{"id": "train_47802", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "ci_integration", "governance", "security_gates"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c48ad8a70e5e6fac32ef73c234a002f2fa85e0e6"}
{"id": "train_47803", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "evaluation_metrics", "auditability"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "666224b83856cc29651ee76820c7a79dc30017d4"}
{"id": "train_47804", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "ci_integration", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a2808dd9b885116b16e1c477b8d4f4ce5a759409"}
{"id": "train_47805", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "cost_latency_tradeoffs", "documentation", "auditability"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f4c87184625c76a4678cdd7b9196b80d1f9e6708"}
{"id": "train_47806", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "evaluation_metrics", "governance", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0d8d76e17a71bfbbb44c4574fcb1d49de00e1a3e"}
{"id": "train_47807", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "security_gates", "documentation", "governance"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9365f13e7de68b932a101a20a700934bd4b1b5b9"}
{"id": "train_47808", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "tooling", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7bc9209a8a922f7ee86145d85670b73f1d070100"}
{"id": "train_47809", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "auditability", "ci_integration", "reproducibility"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3457ab8cb0f90502ca0da41bcbd490c06f0c9d99"}
{"id": "train_47810", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "repo_scale_reasoning", "reproducibility", "tooling"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4b38bf21315dc5aaf15da3705f5d919c48e29603"}
{"id": "train_47811", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "governance", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b6d9b11302c027bc7fe313dd7304d987c75fd33a"}
{"id": "train_47812", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "auditability", "documentation"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7c80dea24bdbba112a4099723436e527f6b6da2d"}
{"id": "train_47813", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "documentation", "auditability", "tooling"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8d293085a3144b9239f48bb8fb6946ef596ed9b2"}
{"id": "train_47814", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "tests_are_truth", "evaluation_metrics", "governance"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6692137605752baba648ae165ade58c1ead6b7c2"}
{"id": "train_47815", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "tooling", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4d8e6a164687cc83ccc1c4b929f9cabc7b26c558"}
{"id": "train_47816", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "evaluation_metrics", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a8467c04bd3693b52c89b60f91f8d4ec0da6abd8"}
{"id": "train_47817", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "security_gates", "auditability", "reproducibility"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2cd58127acb90b2b54cbb7b9a6285455cd817351"}
{"id": "train_47818", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "repo_scale_reasoning", "evaluation_metrics", "reproducibility"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e8469aeaf1cdfd6b81ec0e5a447f7fb72b789702"}
{"id": "train_47819", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "auditability", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "84dc69c88b1b0d29131081ba7559b36fee02e3ef"}
{"id": "train_47820", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "auditability", "tooling", "documentation"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "be5b42323403d1d827fd55925fcbcbe06bac9551"}
{"id": "train_47821", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "ci_integration", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eebb9ee35769d9208a0a07b063be91c3b0dd6ae6"}
{"id": "train_47822", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "documentation", "evaluation_metrics", "governance"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "913ece1a7720b803767b0e1b810b3e996eaa2b6c"}
{"id": "train_47823", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "governance", "reproducibility"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ea11daaec00571dd191b656a1b4555ce1782d2f8"}
{"id": "train_47824", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "security_gates", "governance", "ci_integration"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "88cd211d7cefa9322cf3727027eb1a13d3f8ef13"}
{"id": "train_47825", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "evaluation_metrics", "reproducibility", "documentation"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "118ff4cadc75e1aa501b5f52dec203cdcf512f08"}
{"id": "train_47826", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "ci_integration", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "88cec2b38409d6158130b487f4cf22f25d801533"}
{"id": "train_47827", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c19e5e0399cd3b9decac3bef673a6ce3d6f001af"}
{"id": "train_47828", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "security_gates", "reproducibility"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f0b95fdbb49c8d5fa3211b5dcfc5056a7a243a36"}
{"id": "train_47829", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "documentation", "governance", "ci_integration"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0c347de8a1f7faece6379e3616f5da4636fabad1"}
{"id": "train_47830", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "documentation", "evaluation_metrics", "security_gates"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1024fa3569426561e444de15811b9d0ca258f0b6"}
{"id": "train_47831", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "security_gates", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc59b00ed472f712a53c82eb3d63f6664747c72f"}
{"id": "train_47832", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "tooling", "documentation", "evaluation_metrics"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0919a29c9ba0fedf5b465ddbc6eb0f3917cd8fa6"}
{"id": "train_47833", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "evaluation_metrics", "tooling", "governance"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ab8909e8368bce74bd7fbbcc48f8bb266ea206c5"}
{"id": "train_47834", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "auditability", "ci_integration"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c5cc9a7a29ce447a31b9d6976d13dacbd41281f4"}
{"id": "train_47835", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "reproducibility", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a9b6a5f9c7d4c019a68688162777a2521177a970"}
{"id": "train_47836", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "tests_are_truth", "security_gates", "documentation"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ec2a9acb4aeba7ff5f74f68dbc882b44811e722b"}
{"id": "train_47837", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["governance", "reproducibility", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8a3139fd16430586cfd96540efd903274520e562"}
{"id": "train_47838", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "repo_scale_reasoning", "governance", "tooling"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6f2a9d4574e558bee09e36b201b58563723c9cbf"}
{"id": "train_47839", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "tests_are_truth", "repo_scale_reasoning", "auditability"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e82f016b2c93351b5d76b0416384e8692ce13211"}
{"id": "train_47840", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "tests_are_truth", "reproducibility", "ci_integration"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aed194d6ac2fe723be733b90f5f256ddb77ba60e"}
{"id": "train_47841", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "security_gates", "governance", "auditability"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f39df8ae7f593490ff6bab8e30fd1215f2e5d753"}
{"id": "train_47842", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "tests_are_truth", "security_gates", "documentation"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7aec9963ffee963c226c535df05f76ff7166c9c3"}
{"id": "train_47843", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "governance", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5c492df3f2f7532e3d8201ab1c2d87a3fd23b9b7"}
{"id": "train_47844", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "tooling", "security_gates", "documentation"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e6a0f5788d306ba692ecccabab1860adf68b8ff8"}
{"id": "train_47845", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "security_gates", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ee12e4bc3e46478db62f23e3a8251a3bf9ac4fd4"}
{"id": "train_47846", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "security_gates", "tests_are_truth", "documentation"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fddf5b63fefda6b98239b03d32f11d0e26b13a8d"}
{"id": "train_47847", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "tests_are_truth", "security_gates", "tooling"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c9164a332534fb31d064b309e90248cefc9a50a4"}
{"id": "train_47848", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["governance", "auditability", "ci_integration", "security_gates"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "24983bfc251a6ec44496cf06055852819998e4e2"}
{"id": "train_47849", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "security_gates", "documentation", "evaluation_metrics"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a5761d610e29864c6d20f0eec2ea9957d11a67e0"}
{"id": "train_47850", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "auditability", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ad981a4919ebe1c9e9cf20d96021f979c4a20464"}
{"id": "train_47851", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "repo_scale_reasoning", "documentation"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "db7fd564199f8c9cebd3ed92ea9c00b16a8dba25"}
{"id": "train_47852", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "ci_integration", "security_gates", "auditability"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "04bbaae5822744ee7dfa0372900683bc2b62c077"}
{"id": "train_47853", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "reproducibility", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "162cc83408e6da7706d56ba31e8a0b61d15a431e"}
{"id": "train_47854", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "tests_are_truth", "documentation", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "80d4574d5555eec06d07e0a9c9ce0dedbbb8c1cb"}
{"id": "train_47855", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "auditability", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4bfbb96a30d7a7fd5836221eda7d52a4f959b4b3"}
{"id": "train_47856", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "reproducibility", "ci_integration", "auditability"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e8563d4877b2fd6f6a7bcaea47d846f4e9ee5a15"}
{"id": "train_47857", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "documentation", "evaluation_metrics", "tooling"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c87470c433349090ce329f09f243740102158f09"}
{"id": "train_47858", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "ci_integration", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "82719cc96e389770ac27ea054a22ee5d9e73774d"}
{"id": "train_47859", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "repo_scale_reasoning", "ci_integration", "tooling"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7d6237aad11eebae3b843e2ba72876db3e12e673"}
{"id": "train_47860", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "ci_integration", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "957aa467d36a5c233f5d595761763dbf96bdef0d"}
{"id": "train_47861", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "evaluation_metrics", "tests_are_truth", "documentation"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5fd5f26bb205f3257f16f534d3f765b1d8fc072f"}
{"id": "train_47862", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "tooling", "ci_integration"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "80aa7d0979caa7dffe1abc328b689bdb9547b8d9"}
{"id": "train_47863", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "reproducibility", "ci_integration", "tooling"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1a586fd0b613ee84f9fa119a406f8033bd8c2f32"}
{"id": "train_47864", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "tests_are_truth", "auditability", "tooling"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6ced6b4e85f266dc5ce408990470b5932a17377b"}
{"id": "train_47865", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "ci_integration", "tooling", "governance"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8cd631a326acb77a94738a1dd99ad973664e0a6a"}
{"id": "train_47866", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "repo_scale_reasoning", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "53f83947ef3fbf3125e2df7e14b39d3ec1656f91"}
{"id": "train_47867", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "evaluation_metrics", "repo_scale_reasoning", "governance"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b86e00d235df3caadf01826130ab2b8be160808e"}
{"id": "train_47868", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "cost_latency_tradeoffs", "evaluation_metrics", "governance"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e0076b0318124b95d786d526138bf4ceeb0195f8"}
{"id": "train_47869", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "auditability", "reproducibility", "tooling"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "93ce79aa2e12257f4f442a7338cae3d50644e309"}
{"id": "train_47870", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "governance", "tooling", "auditability"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4e6f602ce5c6188f8ec037eb8327438c827741e7"}
{"id": "train_47871", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "tests_are_truth", "reproducibility", "evaluation_metrics"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8d2bf8fb02349545cb0c446d361ec92d5c7ea336"}
{"id": "train_47872", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "tests_are_truth", "repo_scale_reasoning", "tooling"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "480aa49732080e13b97bb705aae6f27fa66443e3"}
{"id": "train_47873", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "repo_scale_reasoning", "auditability", "tests_are_truth"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e2bcf02d2c096b6337567adc48b26059269c4ceb"}
{"id": "train_47874", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["governance", "security_gates", "tooling", "documentation"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9a4a5048ca10e7709d6df5853b33ca894d2130cc"}
{"id": "train_47875", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "documentation", "security_gates", "repo_scale_reasoning"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2c1fb858fcd83621123ecaf5098a908901eda622"}
{"id": "train_47876", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "tests_are_truth", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5d651c59155f6ab75cdbb8192a963500d4a91120"}
{"id": "train_47877", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "evaluation_metrics", "auditability", "reproducibility"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1155cf217b74f8a0a64b9561d763879209265c2f"}
{"id": "train_47878", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "evaluation_metrics", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ed8d4ec88ca1c275ba6d589cf9ea656298e07e53"}
{"id": "train_47879", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "cost_latency_tradeoffs", "documentation", "security_gates"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dfe6dad3aa3d4ed2dccae6c0e408f0111a08eef8"}
{"id": "train_47880", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "security_gates", "governance", "ci_integration"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4d9c6370c52a9dc1ecdccf393da7dfd7abe9d5f6"}
{"id": "train_47881", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "governance", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9ae256fd5f56ff85a25fdb113dee6ecec015117a"}
{"id": "train_47882", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "tooling", "reproducibility"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "abf8caff00ef9339ab0d1975b0d71472a2ac3e57"}
{"id": "train_47883", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "tests_are_truth", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5c532423fb917e318e7913d2ae3d5a578679253e"}
{"id": "train_47884", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "governance", "tests_are_truth"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "af581c4fd37c5353943d43c3f25a2aac25b07b0e"}
{"id": "train_47885", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "tooling", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8a41cde524e82ec46964b6134592daa045976084"}
{"id": "train_47886", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "reproducibility", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "252f8276b29ee6ce38ee7efcd4856a3f66478e10"}
{"id": "train_47887", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "governance", "reproducibility"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6a8f164d507b57deef34add8962b289ee8f41ea2"}
{"id": "train_47888", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "auditability", "security_gates", "evaluation_metrics"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9b39cdbc92d28963ed66f3dac9aaa06e810d6e80"}
{"id": "train_47889", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "documentation", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c3b84390f20fda74dff4c8954705e31e32e364f9"}
{"id": "train_47890", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "tests_are_truth", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f9e95ab90f8c6eab2caf9812c80ee8655b256a46"}
{"id": "train_47891", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "documentation", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "788e4f444cd6c11b077b39efb6d95060f99f5409"}
{"id": "train_47892", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "security_gates", "tooling", "governance"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "05512582ce99d03d57bed3a70cbcf4d0d13d3322"}
{"id": "train_47893", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "governance", "reproducibility", "documentation"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c29788ad2991b97036f02cde2902e86ea17c096a"}
{"id": "train_47894", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "auditability", "tests_are_truth", "tooling"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "093d523a92a55d6f5041a17761c6b825ca85e1f4"}
{"id": "train_47895", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "repo_scale_reasoning", "ci_integration", "auditability"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4ea35580e409eb99bb5a99e17254c1ea1ed9eb32"}
{"id": "train_47896", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "tooling", "reproducibility", "documentation"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4d849270e865d8cf3096eebc788d260df92b0129"}
{"id": "train_47897", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "auditability", "governance"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ad377efb6052a61c9d0179d28924c30b3a6f6a17"}
{"id": "train_47898", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "documentation", "tooling", "reproducibility"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2ee1e0b444da60df9208d10b361605b76cc573b9"}
{"id": "train_47899", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "tests_are_truth", "reproducibility", "evaluation_metrics"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f4d5bc3bd42eaeffed479ceb8b6fbf725b9f5ae1"}
{"id": "train_47900", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "auditability", "ci_integration", "documentation"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "42ff26ae17345ce63726ac4469edefcc9b9af93d"}
{"id": "train_47901", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "documentation", "auditability", "tests_are_truth"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8b0bad73fc9ad2036fcd41151be1a7017fdd4f1e"}
{"id": "train_47902", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "ci_integration", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "258bb4bfe2736c0d9bb6550d42a83d58694673be"}
{"id": "train_47903", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "evaluation_metrics", "tooling", "auditability"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f7331397720bb7dbc17c017e3135528226997172"}
{"id": "train_47904", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "security_gates", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "04879379981de9dfda0e199b92c238a080b3cde7"}
{"id": "train_47905", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "documentation", "reproducibility"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "caa1417806654445b9216a6dbdf459e6300f6621"}
{"id": "train_47906", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "governance", "evaluation_metrics", "auditability"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "03416376969c0532b5f8ab78fdcc2bbb3fde8301"}
{"id": "train_47907", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "tooling", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "67f0d354f3920b4ccd0bef61122f7fc5a01bc1dc"}
{"id": "train_47908", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "tests_are_truth", "tooling", "documentation"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6197dd75b7938446ade53d5846aba8f1b28c74fe"}
{"id": "train_47909", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "auditability", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ed8d4ec88ca1c275ba6d589cf9ea656298e07e53"}
{"id": "train_47910", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "tests_are_truth", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bdac62d34b33586e6bef83213c0535d14c712a84"}
{"id": "train_47911", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "documentation", "security_gates", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "752a044ac99ed40945fb0b7ffc9c2ca3160f5bc0"}
{"id": "train_47912", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "governance", "auditability"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "afaad93acbc3cb94b230f3922f43c6c58d4bf3aa"}
{"id": "train_47913", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4994cc8a95b0872556ed9420261c7c7e4fca6442"}
{"id": "train_47914", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "documentation", "reproducibility", "evaluation_metrics"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8aa78c467ca84bdb2d3e31208f74895662be4bb2"}
{"id": "train_47915", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "security_gates", "auditability", "governance"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e451a4f3b56baa760b52131d437b6a1bc05bbe36"}
{"id": "train_47916", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "ci_integration", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1c26700d2fb0c4bf4e01fb34e43413efd0c8cd8a"}
{"id": "train_47917", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "security_gates", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d006ac430b5e1d79f0f92222ea12ea077846f78e"}
{"id": "train_47918", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "tooling", "tests_are_truth", "governance"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ccef1d1f626a6fe26eb5bc298aba1fec40084957"}
{"id": "train_47919", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "cost_latency_tradeoffs", "reproducibility", "documentation"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dbfb42d14d843819b5bff66f9f35a69cfaac56c7"}
{"id": "train_47920", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "tests_are_truth", "governance"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "258bce57b732aeb3ca53ebee74c8168a8cdf699d"}
{"id": "train_47921", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "tooling", "auditability"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1e34cca04c25e964cbc0c3701ee8d481f43c66dc"}
{"id": "train_47922", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "evaluation_metrics", "ci_integration", "documentation"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9031a2e18d45b49c529e33cf7325d9e8be89ea37"}
{"id": "train_47923", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "security_gates", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "175bc2ce9dc2135c5d7dfc19440dc0f178d45785"}
{"id": "train_47924", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "tooling", "documentation", "repo_scale_reasoning"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1b6b58981169ceda19e7fd49777343862d4df4c8"}
{"id": "train_47925", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "auditability", "governance"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9db365cad2e0283b4e72d8ab9222e1fe9e37500d"}
{"id": "train_47926", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "reproducibility", "tests_are_truth"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "07b71f88ea4ca89d74fbbd7389703b323f7847be"}
{"id": "train_47927", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "ci_integration", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "365ea6aeea9a99df971366a2b39779703d21e451"}
{"id": "train_47928", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "security_gates", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "762bb4efe9105e3d8b751726b0ecad5a0bab352a"}
{"id": "train_47929", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "tests_are_truth", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "67674507638e4253ade2509ad90b4e0614f90f1f"}
{"id": "train_47930", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "tooling", "documentation", "security_gates"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5a688b549e60e1ee100f43f77ddff67c643e0fc4"}
{"id": "train_47931", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "documentation", "security_gates", "ci_integration"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e910d2c94f57dd63d51d92f7dd486c19a2e50f54"}
{"id": "train_47932", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "tooling", "auditability", "documentation"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4b777104820a39ebdbc2cd9a44cac053d94166f5"}
{"id": "train_47933", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "auditability", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b7db73e10c6ae2686d9888f0f2ce7e078d3b5485"}
{"id": "train_47934", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "tooling", "auditability", "governance"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ee2459421759de5a51206b9f6186b01413f23370"}
{"id": "train_47935", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "security_gates", "repo_scale_reasoning", "documentation"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ba24c042848c2a8bf7a0aed7b71904283e22a2ca"}
{"id": "train_47936", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "tests_are_truth", "security_gates"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fadfcbf18bc8b69a90944f43d6551da69cdec018"}
{"id": "train_47937", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "repo_scale_reasoning", "security_gates", "tooling"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0c23458e4fcb217c4bbef4858c50041a765dd9c5"}
{"id": "train_47938", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "governance", "tooling", "ci_integration"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0eee33d8eb79272c09e16f9559b0daaa1cf9f007"}
{"id": "train_47939", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "documentation", "reproducibility", "governance"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6e31b8c76327a8cdfc0abf58be44fd2111b32055"}
{"id": "train_47940", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "repo_scale_reasoning", "tooling", "governance"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "353bd4ca3e1bbe9f6f22af846bd16b5487b7ff21"}
{"id": "train_47941", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "auditability", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "40512d3b4ba04bce731a49ae664c29305fae2651"}
{"id": "train_47942", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "governance", "reproducibility", "security_gates"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8b299289ecc4cad3a8af417aa7df23e9ee308b8e"}
{"id": "train_47943", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "382a394f7dfb88e9b3f330f20d968435cfce1e9d"}
{"id": "train_47944", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "tooling", "tests_are_truth", "ci_integration"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4414a0bd6017dd9d077f292870bae3a1441fb721"}
{"id": "train_47945", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "governance", "tests_are_truth"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "949a762e72e4a67327797ad24966e0ba8627342e"}
{"id": "train_47946", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "cost_latency_tradeoffs", "reproducibility", "tests_are_truth"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b056ba5c140d0babf88f578dfbd94d691bcf69bf"}
{"id": "train_47947", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "auditability", "evaluation_metrics", "governance"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "37eba4c6a2fe21091db7358967b8c15edbc2c005"}
{"id": "train_47948", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "governance", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fac9e864180915821743acd52cb2d57b96732b48"}
{"id": "train_47949", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "tooling", "evaluation_metrics", "auditability"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c4801de5e3df90476ee6e4dbbad65f9d45d6622f"}
{"id": "train_47950", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "governance", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d160c81e40bb990a876632aa751860372af063b7"}
{"id": "train_47951", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "repo_scale_reasoning", "governance", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a20417c5ae7fd7ba1bb420d8bd91cd0ddffaab19"}
{"id": "train_47952", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "auditability", "tests_are_truth", "documentation"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d3ef700992a90cee47d336786f953b7ae26ffcef"}
{"id": "train_47953", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "evaluation_metrics", "governance", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c64bb68a97dc5c53d9f44faf55ee1773a9099c3e"}
{"id": "train_47954", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "repo_scale_reasoning", "evaluation_metrics", "tooling"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cd3c9f816c41b5e83358a893e7331a6aee2652e7"}
{"id": "train_47955", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "documentation", "tooling", "tests_are_truth"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7b17722a529ef199737bc0cadac3282fbe9e0994"}
{"id": "train_47956", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "35bd1d79b2b8e198e28abb7e2163b290d8021b44"}
{"id": "train_47957", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "auditability", "evaluation_metrics", "documentation"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "30007f7efbbbeb72ced8fbc0f7384bbebe7c94c4"}
{"id": "train_47958", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "auditability", "reproducibility"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dd43757447dabe24fc144eab95ccdfa1a938aa8a"}
{"id": "train_47959", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "security_gates", "ci_integration"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "143294b4747a05dcb794fddd70e34d27239cae8d"}
{"id": "train_47960", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "tests_are_truth", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "73101921169c5ce7b7673407e4e62e000d5db79c"}
{"id": "train_47961", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "security_gates", "governance", "documentation"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1f6de1c84518bb75e129872905b7ba59fa644d3b"}
{"id": "train_47962", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "tests_are_truth", "evaluation_metrics", "reproducibility"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b1258f89daadc25e1e4c6433e002b3f485ded9e0"}
{"id": "train_47963", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "reproducibility", "security_gates", "governance"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eebeb89bfaa6bfb01f6956937e2d8bb167bcd63a"}
{"id": "train_47964", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["governance", "evaluation_metrics", "documentation", "tests_are_truth"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "db69b7b9f1e36eff38c089a8a4db34cc8b07a222"}
{"id": "train_47965", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "tooling", "governance", "tests_are_truth"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "03fb2111c0beefdf3aac5b3dd5e769c4280118cd"}
{"id": "train_47966", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "governance", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ced8577698dea8c61394b3a2c6220067a6257c5a"}
{"id": "train_47967", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "repo_scale_reasoning", "security_gates", "reproducibility"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0ef66801e2391ba6d7a9d0bc1eb93625d38e48c1"}
{"id": "train_47968", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "governance", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e10e5b11cd9d3c5532ebb9b3bdb0ddaf7404a54c"}
{"id": "train_47969", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "auditability", "repo_scale_reasoning", "security_gates"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "421f67df90eb0ebfaf66e9429038a959963d1d7d"}
{"id": "train_47970", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "cost_latency_tradeoffs", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b126a8d0418210478d58df6b8c2add0c0271aae3"}
{"id": "train_47971", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "ci_integration", "documentation", "auditability"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d7eb825c4616757ea5581d76e5ce72eb02c66b3b"}
{"id": "train_47972", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "governance", "tooling", "auditability"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a4fd7eeb252152489a64503e952040f4523d1795"}
{"id": "train_47973", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "ci_integration", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3368801aa74963e193ae2f9648882c4eb3afc9ce"}
{"id": "train_47974", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "governance", "tooling", "security_gates"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "afc494647434a93eae23e35f324087de431457d8"}
{"id": "train_47975", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "security_gates", "tooling", "auditability"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3686409d01a8eb5efb8dfdd0818f089a2c6410f5"}
{"id": "train_47976", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "evaluation_metrics", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "317413e2534f966674f545ae55e2c3e63a9cbc28"}
{"id": "train_47977", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "documentation", "ci_integration", "tooling"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c145c97d613fa80154a457028eac8bcbb50378db"}
{"id": "train_47978", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "ci_integration", "governance", "reproducibility"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7fa36afeb62e9deddd8cccf36ba0def7f317576c"}
{"id": "train_47979", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "tooling", "tests_are_truth", "ci_integration"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b63faf6c23263ba6b928b65efaeae6ea59d69b02"}
{"id": "train_47980", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "tooling", "governance"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "505010755f445bb92a542582640471d1a81c603b"}
{"id": "train_47981", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "documentation", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0291e23388797aa97b7f642754ee430b964bd2df"}
{"id": "train_47982", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "cost_latency_tradeoffs", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b21ae0aeba9b17fde5cace055dea2b09179c2ddf"}
{"id": "train_47983", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "repo_scale_reasoning", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "acab0c0f52789f93190bce4a3096b4b99f78b88d"}
{"id": "train_47984", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "auditability", "governance", "tests_are_truth"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c16b1537e3fff653656886edf5635f2f96f5c272"}
{"id": "train_47985", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "reproducibility", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cc7b499f5a60b7ca224d4051c1e3ab96dcf39045"}
{"id": "train_47986", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "reproducibility", "evaluation_metrics"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f8a47e82babe59f13035a119e38591fda61171dd"}
{"id": "train_47987", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "ci_integration", "governance", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d1e62de3b7371aa23799b339bd52b1606da66954"}
{"id": "train_47988", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "evaluation_metrics", "documentation", "reproducibility"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ced8577698dea8c61394b3a2c6220067a6257c5a"}
{"id": "train_47989", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "evaluation_metrics", "governance", "tooling"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d5e3668e02ea243df5d9214fbe4d782fcd6064f2"}
{"id": "train_47990", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "tests_are_truth", "documentation", "tooling"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9d846669d081406c518a27b9904d0cfcd6dcfbdc"}
{"id": "train_47991", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "documentation", "auditability", "governance"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "03201ba8edfd3a4f75294c80f6ed2736a542c64d"}
{"id": "train_47992", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "auditability", "security_gates", "reproducibility"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "631eebc58e8484db95cba88b3b4ab2a14a0171f4"}
{"id": "train_47993", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "reproducibility", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5aefa14e1f775d6693fcb6a0299b1583a5c82730"}
{"id": "train_47994", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "56aa9f6fc91786642ddcb00d941578180d61fcf1"}
{"id": "train_47995", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "reproducibility", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1ec6bda08136c38d9e74e225ede5112ce7dd556e"}
{"id": "train_47996", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "reproducibility", "tooling", "security_gates"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "09154ec57007bdc229309d280852be8bf769deeb"}
{"id": "train_47997", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "auditability", "ci_integration", "documentation"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8438c4e0f99838b9b8e4a1e1c9740b255dc30d65"}
{"id": "train_47998", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "488a0b3281011bf8533f7ac504b32219d2da5bd4"}
{"id": "train_47999", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "tooling", "governance"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ddcd471e6318b2fcdd152fc7c0b5d8ee84e3ba01"}
{"id": "train_48000", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "repo_scale_reasoning", "governance"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4da3e68651bf56906fa8ab1a28d7f637a901e68d"}
{"id": "train_48001", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "documentation", "tooling", "auditability"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "52bd0ba10ae7b4bc179b42b6c8359d141dab2b4a"}
{"id": "train_48002", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "auditability", "tooling"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c3212e1906fc9b1b00882ab3953c2e5677276910"}
{"id": "train_48003", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "security_gates", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ba5d89071932df315ca511e08a237aa371f5e945"}
{"id": "train_48004", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "documentation", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7228f653b1a31bcfb214373d7f1b4cf7b14eb65e"}
{"id": "train_48005", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "governance", "auditability", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b14c3b0ae8a6d2ca07fc2055d67eefb63e936c37"}
{"id": "train_48006", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "tests_are_truth", "auditability", "reproducibility"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8a794ddf168222bf51bee1856a9e7bab6bd79fed"}
{"id": "train_48007", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "evaluation_metrics", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5a1f7ef37a1122b6582deddc6247a14a4136b47d"}
{"id": "train_48008", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "tests_are_truth", "documentation"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4698c5908a4998dadf5cfcdddfaeb65f23bd42b1"}
{"id": "train_48009", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "evaluation_metrics", "cost_latency_tradeoffs", "governance"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "72d06068c9ea7091caf126cde46fc6afffbdca07"}
{"id": "train_48010", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "tooling", "cost_latency_tradeoffs", "governance"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "83875c8c03d109146fb60731896613adb53228ed"}
{"id": "train_48011", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "ci_integration", "repo_scale_reasoning", "governance"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3af11f24d4d2e9e6e963d078afb74789d84ec61b"}
{"id": "train_48012", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "cost_latency_tradeoffs", "reproducibility", "tooling"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "638707af24184947db0d5d5914f53ba3adfd1ef7"}
{"id": "train_48013", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "documentation", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0da98d1a1f37b4647ed4d639915602a08f50fba7"}
{"id": "train_48014", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "auditability", "tests_are_truth", "security_gates"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8aa78c467ca84bdb2d3e31208f74895662be4bb2"}
{"id": "train_48015", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "ci_integration", "security_gates", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "98722104998b04e2787c752f3fecf346249e31f0"}
{"id": "train_48016", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "ci_integration", "reproducibility", "evaluation_metrics"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ce2feaababceb66173a08eeaa1848ed1f1ae04ac"}
{"id": "train_48017", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "auditability", "tests_are_truth", "tooling"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "36769b3685b0f5c5bc87a01e29dbc1f1e5c5ceed"}
{"id": "train_48018", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "evaluation_metrics", "security_gates", "auditability"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4dd0264d94027a235ed4c65d222af0f299f99571"}
{"id": "train_48019", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "documentation", "tooling", "evaluation_metrics"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b0aa47f4c7362300d754ccf75fea46f7502ac9d4"}
{"id": "train_48020", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "repo_scale_reasoning", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "25143162f230c6780081c708795f96ccd0da37b6"}
{"id": "train_48021", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "tooling", "repo_scale_reasoning", "documentation"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0717bbab8d779be708723acbad2e7ce1bdc87327"}
{"id": "train_48022", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "auditability", "ci_integration", "documentation"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3a852b09abba5f51339c479b8c996ffa96d786df"}
{"id": "train_48023", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "auditability", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "90061088641529ffd9a49b7fb4105dc78e752b26"}
{"id": "train_48024", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "ci_integration", "tooling", "security_gates"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2edd7208d73c69cd0498ed13948982f41c99905f"}
{"id": "train_48025", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "evaluation_metrics", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "148c68bddcaea4d921c8ff883ea94091b7fbc521"}
{"id": "train_48026", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "documentation", "governance"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4532bcd9c8fa71ca73d31fed3ba4fa85403f9025"}
{"id": "train_48027", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "tooling", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "da253274444462a3a56bbf0acabf9ce6a576f445"}
{"id": "train_48028", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "repo_scale_reasoning", "auditability", "documentation"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8c674d767f439b31a8f6738208ec6561cbd63161"}
{"id": "train_48029", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "tests_are_truth", "tooling", "security_gates"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e5c0266463166559aa6f2587526d9a9a06450464"}
{"id": "train_48030", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "evaluation_metrics", "reproducibility"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3061779f9bfd6e4bcb2b60b54b47122d5dd744d7"}
{"id": "train_48031", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "tests_are_truth", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0b841cfc7d03819b30ff70aeff5a6c7ee8ae0249"}
{"id": "train_48032", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "security_gates", "auditability", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5d96090bcd4c37a47ac087ec8bce4593fec45944"}
{"id": "train_48033", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "reproducibility", "auditability", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eed67ba6cb3af11e05409f18b290cc43f95c74be"}
{"id": "train_48034", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "ci_integration", "tests_are_truth", "auditability"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aa89f599f2cbcdf0226cf81de46ddd76a5ae7b4f"}
{"id": "train_48035", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "tooling", "tests_are_truth", "security_gates"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6c4a39adebe7eb5614cc3c5d0bcd56cb8484453c"}
{"id": "train_48036", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "documentation", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6ffdc5ff26d73bcc35c4f21268bbc93b96db0eaf"}
{"id": "train_48037", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "governance", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "65beedf2f61f879c71e5d519410ec6f0cac92ce6"}
{"id": "train_48038", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "ci_integration", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f785220e091cf1c244d216868f172d257938f50f"}
{"id": "train_48039", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "ci_integration", "tooling"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "169a7a4078044c974d82819a937a2c38ee28a9df"}
{"id": "train_48040", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["governance", "ci_integration", "tooling", "reproducibility"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "44fd81ec4ce931ea0cf8b7adf6b85a7172a7f93a"}
{"id": "train_48041", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "tooling", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ccc97562409331af5922aa8f3813287d9d8b577b"}
{"id": "train_48042", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["governance", "ci_integration", "documentation", "evaluation_metrics"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a4b5c0aafebc20830b5578f36b412ffcf02d22fc"}
{"id": "train_48043", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "documentation", "reproducibility", "evaluation_metrics"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f1d9d896058a2fdbfd43d3c9eb0853b3f65cf918"}
{"id": "train_48044", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "security_gates", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "462a7d383e34ddae6a3d76e400b4659ee8967785"}
{"id": "train_48045", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "security_gates", "auditability", "tests_are_truth"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "45d6e0b3ee4641465c2b95a7645f3aa239447aca"}
{"id": "train_48046", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "cost_latency_tradeoffs", "auditability", "documentation"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "64ea2bb53602ccadd9ff600d80a3704457aded73"}
{"id": "train_48047", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "documentation", "reproducibility", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dd4ddcfb8fc061f5d0e468d30257a481069e15e1"}
{"id": "train_48048", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "documentation", "tooling", "ci_integration"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "62750cfaefab5f99033dd1bb38cee07855f4cf56"}
{"id": "train_48049", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "ci_integration", "auditability"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f7b0d99b541592fce004e4bc2cfee5a93d7f12e4"}
{"id": "train_48050", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "reproducibility", "governance"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e579c22ea86dd64a15ac1ef790a5855e69d5c4fa"}
{"id": "train_48051", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "security_gates", "auditability", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7aadf648dc7e729cd50163930bc4cd7b35282497"}
{"id": "train_48052", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "ci_integration", "security_gates", "auditability"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "edf5a108d7e1e0158248e546639e154d7f5d55b3"}
{"id": "train_48053", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "governance", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "35267c1a0c4f564d4d8977a0ee0678d299168c1e"}
{"id": "train_48054", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "auditability", "governance", "security_gates"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7a9783cea5d90177a5b791d64127bdf238d3ff34"}
{"id": "train_48055", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "ci_integration", "documentation", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fd68868ed678ad6f3c6d043bfc28efd60631f2ff"}
{"id": "train_48056", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "auditability", "tooling", "ci_integration"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "30e8a21ec10839b32a979bd9d2052ebd69e41450"}
{"id": "train_48057", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "tooling", "documentation", "reproducibility"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b91a1f501189619c384e03f6569df72277967225"}
{"id": "train_48058", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "auditability", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ca093fc201fe2e1d59f03b5175ac535e24dcfcd8"}
{"id": "train_48059", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["governance", "tests_are_truth", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "270ed0ba27b6f893982c1ce3a4d093570ea0642e"}
{"id": "train_48060", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "reproducibility", "documentation"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "15562bba4c69541e9e5d5f6d36725e7163842959"}
{"id": "train_48061", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "documentation", "auditability", "governance"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "af46df9a16004c19fe3deb42329d62b19c84ab8f"}
{"id": "train_48062", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "reproducibility", "governance", "evaluation_metrics"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a9cbbdaf87875ee733980529a66d25877e0535ac"}
{"id": "train_48063", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "security_gates", "evaluation_metrics", "ci_integration"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b86e00d235df3caadf01826130ab2b8be160808e"}
{"id": "train_48064", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "auditability", "governance", "documentation"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7485fe4aae4370d8ea60dc915b2ce34501d70566"}
{"id": "train_48065", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "tests_are_truth", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f424a75c64f9711a7a0b836a1f5de424d66d9c6d"}
{"id": "train_48066", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "tooling", "documentation"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "815923c6c1981f01265bc3fad5e7ef6fc3513e6f"}
{"id": "train_48067", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "repo_scale_reasoning", "tooling", "evaluation_metrics"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "767ac1923594a1f790bdc7cec8a45d5c39dab5aa"}
{"id": "train_48068", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "repo_scale_reasoning", "ci_integration", "reproducibility"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "86e17ec31babe7bec15c889ed355f8f8fdf5fb2e"}
{"id": "train_48069", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "security_gates", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a98ac9364be59940c33dc5e01987d62c0ac702f8"}
{"id": "train_48070", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "security_gates", "documentation"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cac6686406ca55ae8225d3b438fe53719568998c"}
{"id": "train_48071", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "reproducibility", "governance", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1de4b60f5a3236a66d2f9398ef70ecc9fc7dda06"}
{"id": "train_48072", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "security_gates", "evaluation_metrics", "reproducibility"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "293fd1f09857e73612b114fc7f76789899057c11"}
{"id": "train_48073", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "auditability", "reproducibility"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "08b58c5351a97153d6e77d156a157dca59c94dbd"}
{"id": "train_48074", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "security_gates", "evaluation_metrics", "ci_integration"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "77b1e62ee94107974dd0c3b8c42a6268872e76c2"}
{"id": "train_48075", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "auditability", "governance", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "16aa509b4084d03426f9bf21413ff1c43616820b"}
{"id": "train_48076", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "tooling", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fa0b47c43f0da7aea307e4f68ae3431e18b277f2"}
{"id": "train_48077", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "documentation", "reproducibility", "tooling"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "928001608bf666a2547c4baab7aa58813bec99f2"}
{"id": "train_48078", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "governance", "reproducibility", "documentation"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "627078819912028cec810f60e01bb27ab29be08a"}
{"id": "train_48079", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "security_gates", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "734f9c378a0857223050843810823e170ce9d27d"}
{"id": "train_48080", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "auditability", "governance", "documentation"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d75ddb794870734285aff369595ddaecabccb688"}
{"id": "train_48081", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "governance", "reproducibility", "tests_are_truth"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "79e1465888fc460128f6f909a03cb68c2b19fcd8"}
{"id": "train_48082", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "ci_integration", "documentation"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "de9c6d193bdd6de106c62234b37bda3c565fa005"}
{"id": "train_48083", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "governance", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bbab33a0606364fa30b5f6e4ddf1a3f9abb76a60"}
{"id": "train_48084", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "ci_integration", "reproducibility"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ffb26eae318ebfe5ae53dec789bffe1c057012fd"}
{"id": "train_48085", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "tooling", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5234ceeea4c3ec6106795333409e4eb033dde46a"}
{"id": "train_48086", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "repo_scale_reasoning", "governance", "reproducibility"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "60deb7e1f227ae109d5fae9b0a8427eccc1c5917"}
{"id": "train_48087", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "security_gates", "governance", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "50e71553eec9f643a2a964155b470b62d4dac269"}
{"id": "train_48088", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "repo_scale_reasoning", "documentation", "tests_are_truth"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8b485a1ae5a641ff37f6ce4cc386f07468ff6cac"}
{"id": "train_48089", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "ci_integration", "documentation", "evaluation_metrics"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "35e46768fdbf6e4448f06f91428009e4430b0f42"}
{"id": "train_48090", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "repo_scale_reasoning", "tests_are_truth", "ci_integration"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "61a7ead3504ffe27225c374f9246ec4050cb6826"}
{"id": "train_48091", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "auditability", "security_gates"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f9eea1f7877a03d1428d1af755e8eaf9fca13c26"}
{"id": "train_48092", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "reproducibility", "documentation", "security_gates"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fb2fa21fbb15cb91022cd487caa370e8deff4d4b"}
{"id": "train_48093", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "ci_integration", "tests_are_truth", "tooling"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ecd8823a69361f27b2ff4bb6eaa6d45a94e88354"}
{"id": "train_48094", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "repo_scale_reasoning", "ci_integration", "governance"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "654e23ec3772c9f2459273b4adc9ca70ebc0220f"}
{"id": "train_48095", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "evaluation_metrics", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3cd5b0999f27497962e5e4d23d9682a1e2f84583"}
{"id": "train_48096", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "governance", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "050718c114a687f42dd02bd598675fdafb7ac33d"}
{"id": "train_48097", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "documentation", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0cc4edffaed7ea9781d0163b23e483d04f5f817b"}
{"id": "train_48098", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "documentation", "repo_scale_reasoning"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5aae1084e9c90e87a94dc174a449882a7bb86eee"}
{"id": "train_48099", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "ci_integration", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0e8e31749d19358df370009266c32ebbc55dcbc5"}
{"id": "train_48100", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "auditability", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4cbfe77fd04e38108d7d445c4a53d0554f6d8182"}
{"id": "train_48101", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "auditability", "documentation", "ci_integration"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "170306f94467c6f1223272f7b35942e7d08fe3c0"}
{"id": "train_48102", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "documentation", "auditability", "reproducibility"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0d684b4827bac8f10a4776445d597bb6b93847f2"}
{"id": "train_48103", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "documentation", "governance"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "49ee86bdf0373f9499d84ceebe9e85c3b32f991a"}
{"id": "train_48104", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "security_gates", "reproducibility"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2f29a90ca7ca3241c29d23ac00fb800a8707a9ac"}
{"id": "train_48105", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "security_gates", "cost_latency_tradeoffs", "governance"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5ec09a49b5b9d986c2f059d8294a46423752479e"}
{"id": "train_48106", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "tests_are_truth", "evaluation_metrics", "tooling"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3fda2bd2ee98ae683ba6ff407324a1ea2a7db0cc"}
{"id": "train_48107", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "evaluation_metrics", "documentation", "auditability"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9f40dfaa4f47b3c690dd7314bd60fe76b661845d"}
{"id": "train_48108", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "auditability", "ci_integration"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "44edcca5bdfe6dc0e2f4f9692f89ca58e11764f9"}
{"id": "train_48109", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "tests_are_truth", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "83c5daf192e23cd57d4418e1005e0a9f4042c794"}
{"id": "train_48110", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "security_gates", "repo_scale_reasoning", "governance"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0b331d8ec0e67e26bda0ce22c8a77bc9702fc8bd"}
{"id": "train_48111", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "security_gates", "repo_scale_reasoning", "documentation"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "899a1ba752cbd198e07bcbf87e38fd45e47593b8"}
{"id": "train_48112", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "repo_scale_reasoning", "ci_integration", "reproducibility"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e6088d4fa84b335832be52c9287e494184f3cc26"}
{"id": "train_48113", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "cost_latency_tradeoffs", "reproducibility", "evaluation_metrics"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c0cf8dbba184c70066ddb781bbb700aee5d18610"}
{"id": "train_48114", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "security_gates", "auditability", "ci_integration"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "178d58c7ff7a172c0817cb44d95a3e3e33e526b3"}
{"id": "train_48115", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "ci_integration", "reproducibility", "tooling"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5ebfffceb75d2ecfbb77ed4c2ab530845f6fa452"}
{"id": "train_48116", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "governance", "security_gates", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b21873863e0e52a0cb3b3d0f7a2b9ec0fe10de90"}
{"id": "train_48117", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "evaluation_metrics", "auditability", "tests_are_truth"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c1e48c8766996b5fc2865f207af4bed3ace4b587"}
{"id": "train_48118", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "tests_are_truth", "repo_scale_reasoning", "auditability"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "df52108b142d67fa025de349cb6a28a6f9b84526"}
{"id": "train_48119", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "documentation", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cbd69dff66da5c4ca3bb87beea7b45a854349276"}
{"id": "train_48120", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "documentation", "reproducibility"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "049c22ba1c3e1bb22964a9cf55a21811efe4b911"}
{"id": "train_48121", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "reproducibility", "security_gates", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4d4c3dda36faad73f4b891723ec7182d869ddbb6"}
{"id": "train_48122", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "auditability", "repo_scale_reasoning", "security_gates"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2b27aeb7edd7c91a54d4156624182b44c0f3934e"}
{"id": "train_48123", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["governance", "security_gates", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "016f2e27d541a4b56ed5171fe9fbbc211bd42e98"}
{"id": "train_48124", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "security_gates", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c592d178d54f58e2d00381a7385094df7e101209"}
{"id": "train_48125", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "governance", "reproducibility", "security_gates"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "35e6a216b43b3b95d8577579bac06fc462c03860"}
{"id": "train_48126", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "auditability", "governance", "evaluation_metrics"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "03bdf5642fa83e14e8fa966001fadd9a6e29ea21"}
{"id": "train_48127", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "auditability", "ci_integration", "security_gates"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "012a08375d3c8eb6889bfaf99403be6539e638e6"}
{"id": "train_48128", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "governance", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f046dad3739a8a28363fc589499c41a8c4fe190f"}
{"id": "train_48129", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "evaluation_metrics", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d28e764ccf05ef268d155f22245ee6a3a0471b1d"}
{"id": "train_48130", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "documentation", "tests_are_truth", "ci_integration"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "140e86b4092dafc1871811fe0763056d5b8ce605"}
{"id": "train_48131", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "governance", "ci_integration", "reproducibility"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "85face5859117e4e01e2bf29116e672266959c5c"}
{"id": "train_48132", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "reproducibility", "cost_latency_tradeoffs", "governance"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "18ca13f5d25697af6402c8147ac6f9f3c39d637b"}
{"id": "train_48133", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "security_gates", "tests_are_truth", "documentation"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "21b92a11c7f74c352478e5cac3fb3c9a4d8ea301"}
{"id": "train_48134", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "tooling", "documentation", "ci_integration"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9d57c87800b1f74ca085e5c485d2b450d5a609e6"}
{"id": "train_48135", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "tooling", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d1dbff6846b77be6290edc577b6ab23c96beb105"}
{"id": "train_48136", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "reproducibility", "repo_scale_reasoning", "tooling"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7745199450edee0aef1440942c53d5c3e17d6bbb"}
{"id": "train_48137", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "governance", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4f24181192847160766880f428b4337324c3efb7"}
{"id": "train_48138", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "documentation", "security_gates", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9005785fc56dda6557ac3d4b68798b9ba07e98b8"}
{"id": "train_48139", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "auditability", "tooling"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "febb7ef05b611dcd8a7d3a9ea535b4295254b402"}
{"id": "train_48140", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "documentation", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "54d5b1460bc596f8fb801ccff71835c478acac0c"}
{"id": "train_48141", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "governance", "evaluation_metrics", "auditability"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3ce23e54c276bc7aec49fe40f4ae61bc298c9d7f"}
{"id": "train_48142", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "governance", "evaluation_metrics", "documentation"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "47f4168c6b1d2e4d5244d6237b391db3d9e49ca5"}
{"id": "train_48143", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "evaluation_metrics", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9092493249bf5ad15d4efbc2bba2982a55c54031"}
{"id": "train_48144", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "tooling", "evaluation_metrics", "documentation"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c9f65ddbccd1f41c5ac832ef781358c8f0bc5a52"}
{"id": "train_48145", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "evaluation_metrics", "ci_integration", "documentation"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c5ecdacb723248c5b44c915149f24a3f9e6ff5d5"}
{"id": "train_48146", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "tooling", "auditability", "governance"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4032351d0072bf16bf14b22270144b522d32b0e5"}
{"id": "train_48147", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "repo_scale_reasoning", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "adb076d40b47518fa12524def3fd82928d93eafd"}
{"id": "train_48148", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "auditability", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c0b5600d136f9b6f4d9cfadbad9b7db842111785"}
{"id": "train_48149", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "security_gates", "documentation", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cd5533e576b501889a3a1de5eaf541f04ea23ff4"}
{"id": "train_48150", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "ci_integration", "governance", "tests_are_truth"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0f658bd4b28c5bd271c9fd6e372abb55ba2eefaf"}
{"id": "train_48151", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "evaluation_metrics", "cost_latency_tradeoffs", "governance"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4d016895a29bddd974c985bb100abed2ba3e75e0"}
{"id": "train_48152", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "security_gates", "tests_are_truth", "ci_integration"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "319c81a95e65c461df7ea1a805e8f16fe2459cff"}
{"id": "train_48153", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "tests_are_truth", "ci_integration", "documentation"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8052f26f2c0cc4af741aabc6dc902644cfb04e45"}
{"id": "train_48154", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "governance", "repo_scale_reasoning", "documentation"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2ac2512f8ef10bd9e997b653aa18b3a5fc38d60d"}
{"id": "train_48155", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "security_gates", "evaluation_metrics", "tooling"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "48b7e6f032864df03634e8e4100ae90f5c91df2e"}
{"id": "train_48156", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "tests_are_truth", "tooling", "reproducibility"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "46acb660a144f8092a17ea0aa0f8d63bc75e40d8"}
{"id": "train_48157", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "evaluation_metrics", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f2fad0768c836e2399dcf11812afd53295236829"}
{"id": "train_48158", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "documentation", "auditability", "governance"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6781c8e633032869a91e3eeac7b86f6129e14136"}
{"id": "train_48159", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "governance", "ci_integration", "documentation"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "548367debbff24708a28e56e87b1be2cf928bdac"}
{"id": "train_48160", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "governance", "documentation", "security_gates"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "18c19e717f037439136bc2a8d8eaf9437e99e0da"}
{"id": "train_48161", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "governance", "auditability", "documentation"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "42703b9cc10b7f05a3062c758bb1141544c53331"}
{"id": "train_48162", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "repo_scale_reasoning", "ci_integration", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5be96790ad0244b658f8853221130c46a7c26c73"}
{"id": "train_48163", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "tests_are_truth", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "52710a70e929185d0ac959009a29a67d577d0549"}
{"id": "train_48164", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "governance", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cbe1061dd28b08412a7f2bfd3ebda7c60da577ca"}
{"id": "train_48165", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "reproducibility", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cf442aa59af6370d0997f951a675ac717f125407"}
{"id": "train_48166", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "auditability", "tests_are_truth", "reproducibility"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc207732f6e7a96f89c39609bb44f0529fb0b0c3"}
{"id": "train_48167", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "ci_integration", "tooling"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fea30f83418c8ef97605968f9f84a22a1ae86f8e"}
{"id": "train_48168", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "documentation", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c98fe176a7ddb044fbc613675cee5064e2095746"}
{"id": "train_48169", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "tests_are_truth", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "125184c89cc4db268c2353bfa5ddbf39806c6a1c"}
{"id": "train_48170", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "ci_integration", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7201f24c603e1f711de8f81e9383e96ca542d9b0"}
{"id": "train_48171", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "documentation", "auditability", "governance"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "85eee64187c76d98c040dcac2b24ace28ded6bd3"}
{"id": "train_48172", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "ci_integration", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "53c337c5588e02e4e5af0a26f4ab6a540c4d4258"}
{"id": "train_48173", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "tests_are_truth", "reproducibility", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7feb4672611480eb7d18054fa4b39c57bb9de41c"}
{"id": "train_48174", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "evaluation_metrics", "documentation", "repo_scale_reasoning"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d65634d6915c018e6b26a32d744e21c6a67b6037"}
{"id": "train_48175", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "ci_integration", "governance"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "82061c3b512ad5604af7185008ed8b85cfe7e12e"}
{"id": "train_48176", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "auditability", "governance", "ci_integration"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "11a8cd245787524c002c11e54c3df723d747eb7a"}
{"id": "train_48177", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "tests_are_truth", "auditability"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "674c2ee41bcb75ccc46702d4d3870c06567aff5b"}
{"id": "train_48178", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "tooling", "governance", "tests_are_truth"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc62b2e8b6593948ea77e4836298fac98176b4d1"}
{"id": "train_48179", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "repo_scale_reasoning", "auditability", "governance"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d89c78acc59f26a0559c5d7ac291d32e9af9667d"}
{"id": "train_48180", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "auditability", "tooling", "reproducibility"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b608634be04a17337913f4f80f5aa505b9d0432a"}
{"id": "train_48181", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "auditability", "tooling"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ee3538508ae08fd27844d4e8a1d1f47b94a3b005"}
{"id": "train_48182", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "tests_are_truth", "repo_scale_reasoning", "tooling"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e7662e4f409034866fcd990e5d192ebf3ad2cfb8"}
{"id": "train_48183", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "ci_integration", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b7aa6e4c72b46fde7802a5913a1340fba147be46"}
{"id": "train_48184", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "tooling", "governance", "auditability"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8414c469959bb5adaf723dc682ef0d0949297b36"}
{"id": "train_48185", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "tests_are_truth", "documentation", "reproducibility"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7d249854c3351920e816cbd8ddeba1c2315932c1"}
{"id": "train_48186", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "governance", "evaluation_metrics", "reproducibility"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d322738eeec620e2b0935340a0766556197c1d02"}
{"id": "train_48187", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "ci_integration", "security_gates"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "41581516a503c00b994fc151f71dcd1b04dc20f2"}
{"id": "train_48188", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "auditability", "security_gates", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e7d3c6012487f81caaab4389eeea1a675406895f"}
{"id": "train_48189", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["governance", "cost_latency_tradeoffs", "reproducibility", "security_gates"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1589e5c42b5c32f1b8f28ff7086c8b4d30e2592c"}
{"id": "train_48190", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "auditability", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "93ffdce60c745ea9124d6fd022ffa381b1fdf2ae"}
{"id": "train_48191", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "documentation", "security_gates", "governance"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3ea22a60edd999adff1a0e9d1618a9eb3e18e517"}
{"id": "train_48192", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "tooling", "reproducibility", "evaluation_metrics"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a97488409a3194c4ec682209301122ee330d1903"}
{"id": "train_48193", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "security_gates", "reproducibility", "tests_are_truth"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5aae1084e9c90e87a94dc174a449882a7bb86eee"}
{"id": "train_48194", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "governance", "tooling", "tests_are_truth"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f7b959ae915dc715c3a54fba8153594a94fb453d"}
{"id": "train_48195", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "tooling", "documentation", "security_gates"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c57564d49c95dcf5dd8de5aa462eb7e3b0330c45"}
{"id": "train_48196", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "repo_scale_reasoning", "reproducibility", "ci_integration"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "318832f9113d9f8263c51f69095129072facf0bb"}
{"id": "train_48197", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "tests_are_truth", "reproducibility", "evaluation_metrics"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "05319f350c172627fad3ac2f20193dc8f1d8cc1e"}
{"id": "train_48198", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "tooling", "auditability"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "edbb3313cef48aaf3789707fa8283faecefa4db9"}
{"id": "train_48199", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "ci_integration", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a4e4c852428f688db3c8100c09cf100bfe4aea47"}
{"id": "train_48200", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "governance", "reproducibility"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "194efddf43335bf4a865f3621814d86b3861f908"}
{"id": "train_48201", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "cost_latency_tradeoffs", "evaluation_metrics", "ci_integration"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "34ce9295466ad942e6bdd808df2b71214c535c87"}
{"id": "train_48202", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "ci_integration", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "74362b8f3864c41838c1a7da8dbc00bb2c0ec5da"}
{"id": "train_48203", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "reproducibility", "auditability"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e70d0af7fbd5fdbf0466acaa4a5a4b03a579d178"}
{"id": "train_48204", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "security_gates", "tests_are_truth", "auditability"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a7bf04507d7b3c1fceb75eeb83bbe70621e92364"}
{"id": "train_48205", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "tooling", "ci_integration"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "195fc948985b283f9fb3821ca83d437b95b6839d"}
{"id": "train_48206", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "governance", "reproducibility", "security_gates"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1d8c147951f59ddd206404bd8749f7db297c2fa8"}
{"id": "train_48207", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "repo_scale_reasoning", "evaluation_metrics", "ci_integration"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "24a9076a7cbbe5f7f892f4e0a67a1be37b1f569f"}
{"id": "train_48208", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "governance", "evaluation_metrics", "ci_integration"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6bcdfc1d58fe6db1321835baaf7b22df9ed1ebcd"}
{"id": "train_48209", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "31717dde7d7dbcdc0c2d2d6013c4fbfab9159d83"}
{"id": "train_48210", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "ci_integration", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "80008ae3ee0e34f176c438e4f90f7ce6cbae003e"}
{"id": "train_48211", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "auditability", "reproducibility", "tests_are_truth"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c1e34154400abe6f945c7361863c72031ad2c033"}
{"id": "train_48212", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "tests_are_truth", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6ab7445d89527e3dd7e358de08f75ca721bd7a54"}
{"id": "train_48213", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "reproducibility", "evaluation_metrics", "documentation"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "78af0edfb43036c00b7c428a9b464083a1035243"}
{"id": "train_48214", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "governance", "documentation", "evaluation_metrics"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "98b5432bc7b7064c41e457193e5d9d0a50e3356d"}
{"id": "train_48215", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "security_gates", "evaluation_metrics", "ci_integration"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0643683fafb87f11298457b5b759bd870ef0aff7"}
{"id": "train_48216", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "documentation", "ci_integration"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1e732f0c13d875c3a0da0240f6a138751be2ee1a"}
{"id": "train_48217", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "reproducibility", "auditability", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5a0614a25ed381745bf60de73d0d2e771aa76f9c"}
{"id": "train_48218", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "governance", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0e7368d83532c0f83c1b210f3b7ab4b02b96ab55"}
{"id": "train_48219", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "governance", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7ddf9e298df2bfa7beb25f956916504fe5e38256"}
{"id": "train_48220", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "tooling", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d7dd00fce40f8dd36e9d60556df393056291f903"}
{"id": "train_48221", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "documentation", "repo_scale_reasoning", "auditability"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f530d27769e933015ffc59e63678877d64298106"}
{"id": "train_48222", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "documentation", "security_gates", "governance"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "df27f4657e34dfacc0474e5b954aa35a69285cce"}
{"id": "train_48223", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "documentation", "security_gates", "tooling"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fb7896b3c4e8b0ccf2a2470e64d939ddb48dd347"}
{"id": "train_48224", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "repo_scale_reasoning", "tests_are_truth", "ci_integration"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2e60b7c25c33bb805c2d5cdf0a135f43e3fea1ff"}
{"id": "train_48225", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "security_gates", "governance", "tooling"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f58fe1b6ce0653e4dcf81c444e15d46a92f97960"}
{"id": "train_48226", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "documentation", "repo_scale_reasoning"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8c0b2bcfc388e21bbe55bc67a4882d96266d1b79"}
{"id": "train_48227", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c9246ff92059bf93a058cf4444404d221b5b7435"}
{"id": "train_48228", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1e88ea99e21fdd479dba6d1f1d41827c03347743"}
{"id": "train_48229", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "tests_are_truth", "auditability"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c91909099a4157549b0cca6d5a4aaa9367dda2ac"}
{"id": "train_48230", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "security_gates", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c6e57866e027fa3718aa46848036157b4eca66df"}
{"id": "train_48231", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "tooling", "security_gates", "ci_integration"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5cff20470d7a6524be4539d9f1f1f99e3ab9de56"}
{"id": "train_48232", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "governance", "security_gates", "tests_are_truth"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0cfc5b16a2b22dc032562cfcf80a9cd75880ee5d"}
{"id": "train_48233", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "documentation", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "70765aa0c0f677a962b90deb950f0aea694a59d1"}
{"id": "train_48234", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "tooling", "auditability", "security_gates"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "975b43f19a2a489f2cbe6105bc61e3ae7edc0c39"}
{"id": "train_48235", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "documentation", "tooling", "repo_scale_reasoning"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9e8340b7c056f9fb55a075b9ed42859c5e966591"}
{"id": "train_48236", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "reproducibility", "ci_integration", "documentation"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dcdc55bbd7fb13d1dd42175d38eec855181f5cb5"}
{"id": "train_48237", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "reproducibility", "security_gates", "auditability"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "456d4467750e2455e0504ad0cd667fa6623a2dd3"}
{"id": "train_48238", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "tooling", "tests_are_truth", "security_gates"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0606ab52886c8b53f0d0b63bdd80ebb25a810ec4"}
{"id": "train_48239", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "reproducibility", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c82933ed06c129fec2e04b874f85d74c35fe33e1"}
{"id": "train_48240", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "758b24eea8eaa569fcae36cb9dcd0b7bc7ffd7cd"}
{"id": "train_48241", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "security_gates", "governance", "tooling"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ca8eaeaf346e029b46c1de9edc6354e329b7df7d"}
{"id": "train_48242", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "tests_are_truth", "security_gates", "auditability"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "66cdd52c9a4a51f169cb488f51834214af7f342e"}
{"id": "train_48243", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "evaluation_metrics", "tests_are_truth", "security_gates"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "51f0e6b16be5e50b4ea4314838a3e10713dc97ef"}
{"id": "train_48244", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "documentation", "auditability"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "917a3c6c12f967dcd1848b24055095a16e5ad604"}
{"id": "train_48245", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "evaluation_metrics", "governance"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ba2ef0a1f28407dc1bdc999d350550d803946bc6"}
{"id": "train_48246", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "ci_integration", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "965e75bb32d592a9cebb3f75017dbf1a34fec263"}
{"id": "train_48247", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "tests_are_truth", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "66a655ed1d62463a1b1d9c80ea2b2cd21625ee84"}
{"id": "train_48248", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "auditability", "tooling", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "049594df47c84809364b96b777dd2d9ad7de460d"}
{"id": "train_48249", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "evaluation_metrics", "governance", "documentation"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d4b58078bdbb01be2de220d9b255377d7ac58ae1"}
{"id": "train_48250", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "tooling", "tests_are_truth", "governance"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "592d0b452c06848a29b7cfb24a3fd1cdaa9261cf"}
{"id": "train_48251", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "repo_scale_reasoning", "tooling", "ci_integration"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bec76fc1c6dcc0d59aa884d417f23f6041de66f5"}
{"id": "train_48252", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "documentation", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "67a4748e8a39ea4796abbcbd8fd0e9f3aa247cc0"}
{"id": "train_48253", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "tests_are_truth", "documentation", "reproducibility"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2374b84bec676f7428b29c648981788787251ff6"}
{"id": "train_48254", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "auditability", "tests_are_truth"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "34c25e5e463d9ba480c71c48a561179c8a13df8c"}
{"id": "train_48255", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["governance", "repo_scale_reasoning", "reproducibility", "documentation"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c1e34154400abe6f945c7361863c72031ad2c033"}
{"id": "train_48256", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "repo_scale_reasoning", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "03ef7abe2fe5dd17ee94d77e61e177d3220594a5"}
{"id": "train_48257", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "ci_integration", "reproducibility", "evaluation_metrics"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0f0356266fe236a1a6f25e7387683eeaf62007da"}
{"id": "train_48258", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "evaluation_metrics", "security_gates", "governance"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f0ae334f216fd2d603ebbac8d83d78d8689cb942"}
{"id": "train_48259", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "security_gates", "reproducibility", "governance"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e5ab96165991ebc7092e5d6dfd79d019616eaa15"}
{"id": "train_48260", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "tests_are_truth", "tooling", "evaluation_metrics"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc9249855c1b710bc9b40c7702fba278ef2053a6"}
{"id": "train_48261", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "reproducibility", "auditability", "governance"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "67334bacb4fe9e9719a26e55e489991ed645f3b3"}
{"id": "train_48262", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "evaluation_metrics", "security_gates", "tests_are_truth"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "55949926a17e901f712b595e8cca73b91722f6e4"}
{"id": "train_48263", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "documentation", "governance", "security_gates"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dd3ae17605b0c17608667cf356c3f7c97d44cc08"}
{"id": "train_48264", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "auditability", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4ec5d26aadb0539b2ba3f3304158315418d2ccf4"}
{"id": "train_48265", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "security_gates", "reproducibility", "governance"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "af813ccd590c2c4b5a176fbde85765c30cd02593"}
{"id": "train_48266", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "security_gates", "ci_integration", "tooling"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "80c9cd241f1e3e86653e5dd2f1599ecc7a47b1b2"}
{"id": "train_48267", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "tooling", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e74be75ce9f1fd9b62c708c5269947aaa4c8cec4"}
{"id": "train_48268", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "auditability", "documentation", "reproducibility"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "18b909ed6a239da303e79e538b7915e5512e957c"}
{"id": "train_48269", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "documentation", "tooling", "ci_integration"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "376916fee74317ad9bbdcb48153fad5df3314f31"}
{"id": "train_48270", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "cost_latency_tradeoffs", "evaluation_metrics", "ci_integration"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3da8fb281cbc6a34d2458f06c76633b6393d17ab"}
{"id": "train_48271", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "tests_are_truth", "security_gates", "ci_integration"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a613e2663361b3469415cada2bd2ae0b7f279b85"}
{"id": "train_48272", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "evaluation_metrics", "repo_scale_reasoning", "tooling"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d9ba0e633bf0768c6888e54d337c0d161552e0a3"}
{"id": "train_48273", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "reproducibility", "ci_integration", "governance"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7095dcdbd18c49bdbf1983dde81790fc12f87151"}
{"id": "train_48274", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "documentation", "tooling"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aa9b28837c311e5fd00e6b3ac3ed51adf6520405"}
{"id": "train_48275", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "security_gates", "evaluation_metrics"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ff4ae34cfc8463732f4fd97b6130a3ce537f676e"}
{"id": "train_48276", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "documentation", "reproducibility", "auditability"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5bb26c477fedd75e28506ef74d35b5e1e3f4926d"}
{"id": "train_48277", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "documentation", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e4ef3606ca615d82a0aef358bb1f81e7f841efaf"}
{"id": "train_48278", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "repo_scale_reasoning", "auditability", "governance"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a59e04380cefe441e7405f1392ae8a3f4134ea7e"}
{"id": "train_48279", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "auditability", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a2400b742dbdee57a2fb4c4be64ce9e14f0ecaf9"}
{"id": "train_48280", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "ci_integration", "documentation", "security_gates"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e3e545e17779e715ac43b71dce543dbcbf8d17bf"}
{"id": "train_48281", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "auditability", "tooling", "evaluation_metrics"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6268cc4f1d8cae312b2534aeb426b5dd24bc434e"}
{"id": "train_48282", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "documentation", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d32dd4fa3d4aaff3d08e5a3c875806222f99f782"}
{"id": "train_48283", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "ci_integration", "governance", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c5ac42c559ff5dad4f674e15752f01a119dd9da2"}
{"id": "train_48284", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "evaluation_metrics", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5c57ad020bba84876ad15164e04b58aa0220b66a"}
{"id": "train_48285", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "reproducibility", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f8e6b02933e390c8c06807d1061d478fa0435fca"}
{"id": "train_48286", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "documentation", "governance", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3b2ee0b250cd8972ac8d014bf5f40d89172ba1f2"}
{"id": "train_48287", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "governance", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dc632641d46e21aee611014b190f658c816f94db"}
{"id": "train_48288", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "reproducibility", "documentation", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "031059f6f20430d1b82b45fa4f8d81187853078e"}
{"id": "train_48289", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "documentation", "governance", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7d23b303c774a7b998be5474623cc168b07f0022"}
{"id": "train_48290", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "auditability", "security_gates", "documentation"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e8d9cb1da200d672fa508e3df6ff66beba9f8968"}
{"id": "train_48291", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "tooling", "repo_scale_reasoning", "documentation"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f169b906d4293d1aa4fe36a931abf71780eb965d"}
{"id": "train_48292", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "documentation", "ci_integration", "tooling"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5b16b1206e2ff34af52d76cc1796f90968e74f42"}
{"id": "train_48293", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "documentation", "security_gates"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d42f8993f8a9471815a6cf11dc9a65c852450cc8"}
{"id": "train_48294", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "tests_are_truth", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ce69a1bb940fe885194ab96fd71dbd4135030ea3"}
{"id": "train_48295", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "documentation", "auditability", "governance"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e5c02c20bf2033133a7912156ee99a9ec45916eb"}
{"id": "train_48296", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "security_gates", "auditability"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2d7c01474ab1910eab9082d7233859f9f3b19de8"}
{"id": "train_48297", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "evaluation_metrics", "auditability", "documentation"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "44edcca5bdfe6dc0e2f4f9692f89ca58e11764f9"}
{"id": "train_48298", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "security_gates", "tests_are_truth", "tooling"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4d67de2d229fccfcde35a90cbd2a2c2d0ebabf0d"}
{"id": "train_48299", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "tooling", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "78603d76027931318ba5ba5931653827fcb35cde"}
{"id": "train_48300", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "tooling", "security_gates", "reproducibility"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4e14b28c6d5f73081f090eb8e600eb06fe79a9d0"}
{"id": "train_48301", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "governance", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b782516573836832a46dc84d630b4772bcba9b2b"}
{"id": "train_48302", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "tooling", "auditability", "evaluation_metrics"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dbca21f4a692bcfe7ad8deae23b4166ed100ed68"}
{"id": "train_48303", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "cost_latency_tradeoffs", "reproducibility", "tooling"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "230be66c434879b7c9bed04b4d6e1464a94154e2"}
{"id": "train_48304", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "documentation", "reproducibility", "governance"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc39ecef4ee90ee69cf3590e37979a5166f7ec5e"}
{"id": "train_48305", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "governance", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9727471b250d74ca27e0add290b4798b23095b92"}
{"id": "train_48306", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "tooling", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bb88abaaa4573e3e3bac719ac61ef424f6249f6d"}
{"id": "train_48307", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "governance", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "656261382a4e38a05373fb7525f28d1b1a25416d"}
{"id": "train_48308", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "cost_latency_tradeoffs", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ad5fb4c5c22bf81347684b7d1f4ceb3d7b5c0432"}
{"id": "train_48309", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "auditability", "security_gates", "governance"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3ec0686284865a507cc1785e8bccd7313dbbc65b"}
{"id": "train_48310", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "tooling", "documentation"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "be62511b2fe1b17704834cfd463e96c5e536fd8b"}
{"id": "train_48311", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "tests_are_truth", "evaluation_metrics", "governance"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f27bc74bec535b1318c77c05bde5e757cef0cdd2"}
{"id": "train_48312", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "reproducibility", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b57c89c3990811fbce53348e889e7c736556deb5"}
{"id": "train_48313", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "reproducibility", "documentation", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f24b524f20fd1b6cffa5419ff3519d7c7bef6faf"}
{"id": "train_48314", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "reproducibility", "documentation", "tooling"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7cdb2ef3c45d31533c2b2eaefbfbd2b1c1baaac4"}
{"id": "train_48315", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "tooling", "reproducibility"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "702dbf27f2f2325b62e727804893f526222944ee"}
{"id": "train_48316", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "evaluation_metrics", "repo_scale_reasoning", "security_gates"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3f20dac38726006c3687cc2e185acedd6280e9fa"}
{"id": "train_48317", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "reproducibility", "evaluation_metrics", "tooling"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "299b4b81948b2db0d72b90f98c6d022a0e01240c"}
{"id": "train_48318", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "reproducibility", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8aa6e451250ca9e2cc397b7ec95cb37a6809d848"}
{"id": "train_48319", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "auditability", "governance", "security_gates"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5c01f47e538459b022ce1d58667af79257adfcfa"}
{"id": "train_48320", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "tooling", "security_gates", "documentation"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3a2bed0db0f63b10a4959893a4d60ab2cde24836"}
{"id": "train_48321", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "security_gates", "repo_scale_reasoning", "auditability"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d5ecbb81a5f9a9856fb0b2b49a5f539631274359"}
{"id": "train_48322", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "reproducibility", "documentation"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc308d43addc17a93b804f0ec1fb3c26ec23450b"}
{"id": "train_48323", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "repo_scale_reasoning", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "68df4ff8a09d9b3d2866aec98c58b1a51b9b81e3"}
{"id": "train_48324", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "evaluation_metrics", "security_gates", "reproducibility"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a1c926ea1a1fdc13feb9fbbe3d42d4289578efc1"}
{"id": "train_48325", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "documentation", "reproducibility", "security_gates"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4d947711df35c9d0c4fb0d68c753f2040d4e9135"}
{"id": "train_48326", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "security_gates", "repo_scale_reasoning", "documentation"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c1099d451e8958a3fbdffe50de9c7088a0df86c0"}
{"id": "train_48327", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "tests_are_truth", "tooling", "ci_integration"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a490765279bb9f6a4f9d63e5d28b021052cad560"}
{"id": "train_48328", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "tests_are_truth", "documentation", "tooling"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5af587e52e4db9b2d92b41caae633440b6c5ff65"}
{"id": "train_48329", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "reproducibility", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2aa1fc4d151590a18105efe76c63960fbb4071cb"}
{"id": "train_48330", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "governance", "tests_are_truth", "auditability"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "215b0a8e0393adbbda41529b4f09ce711862686a"}
{"id": "train_48331", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "security_gates", "evaluation_metrics", "ci_integration"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8d2bdab8259e529aa28276704601e47f9a31c65c"}
{"id": "train_48332", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "evaluation_metrics", "tooling", "ci_integration"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "af1f789b0d4b82ee595753d8583c77e656cc6c96"}
{"id": "train_48333", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "evaluation_metrics", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eb54ef4a3198491078d12b257eadae67f68ae3a6"}
{"id": "train_48334", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "security_gates", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f1d9c9587ba11a28b20cea8d432074c1015803c1"}
{"id": "train_48335", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "documentation", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c45fbb2b432e148f0f94c3cf93c4bcd587ad86c0"}
{"id": "train_48336", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "cost_latency_tradeoffs", "security_gates", "tests_are_truth"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bada1107e109aa1d3fbea9398af7b853aee4c038"}
{"id": "train_48337", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "reproducibility", "auditability", "documentation"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d557e21fe235028d0e707cdf3ae8c17e1bc7feb5"}
{"id": "train_48338", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "tooling", "repo_scale_reasoning", "documentation"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0250b13089db42cf8d95c550a8bc338df9702e65"}
{"id": "train_48339", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "auditability", "security_gates", "documentation"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7eca5bd768663ef2f408d44d60de8d0e18eac0e2"}
{"id": "train_48340", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "security_gates", "tooling"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d5c5f7ae78445ed0b13a84cb6c38658a4b786883"}
{"id": "train_48341", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "governance", "documentation", "auditability"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ac269f005962e9fedea0e81fc96076e204cb9626"}
{"id": "train_48342", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "repo_scale_reasoning", "ci_integration", "tooling"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0bc406ecf7864e64ba99bf37565d3936cc0cca8d"}
{"id": "train_48343", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5db2320ce4a14193f1599201e97fca32aa8fbba1"}
{"id": "train_48344", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "auditability", "ci_integration", "tooling"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0392945190ca56b272d05d5bbce244f1774e2dda"}
{"id": "train_48345", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "reproducibility", "repo_scale_reasoning", "documentation"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "78da504c7d0c219674bb80e114cbdfb3919b6575"}
{"id": "train_48346", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "ci_integration", "documentation", "governance"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1c455ab04c617100aef869aa5bb7a6aad315cc80"}
{"id": "train_48347", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "security_gates", "ci_integration", "tooling"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "48b2fe23193c1dcd2199e5f82afc1977f93c864a"}
{"id": "train_48348", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "tooling", "auditability", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6b5693b425d9b1d1e1c53ac8c573b76863b4bc35"}
{"id": "train_48349", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "governance", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6e04fda6897b9e31c7810ee017e71a50f1e20e4e"}
{"id": "train_48350", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "reproducibility", "ci_integration"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "039b71644af2d8cd8226f469a05c79d4d6d03fa8"}
{"id": "train_48351", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "ci_integration", "security_gates", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "06dc65d1e1cacccc00f52533b9bb6937c53ffef9"}
{"id": "train_48352", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "security_gates", "tests_are_truth"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e04c4e3042c851f1f11eb4f2aa1b1c974799982a"}
{"id": "train_48353", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "security_gates", "tooling"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "443bc8b078b21e36c0ed469f0844d7af3a09ed08"}
{"id": "train_48354", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["governance", "documentation", "reproducibility", "security_gates"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "075299a03394a92f53b340537081191de8ce17b0"}
{"id": "train_48355", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f8ef4c9c7e4bc62f0b2a812850acaa1482398cd5"}
{"id": "train_48356", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "evaluation_metrics", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b2e4c4953e4c2ac7282b284ce434beb85cf61c3b"}
{"id": "train_48357", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "ci_integration", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "26d189db008b20c80f61bfc37e2897e27f863d49"}
{"id": "train_48358", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "governance", "evaluation_metrics", "ci_integration"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f13431bd7baca643d3dfd392700ff1462361cfca"}
{"id": "train_48359", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "tooling", "documentation", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "df4b422910865bb28b75a3dc302ca27e3ebbef3c"}
{"id": "train_48360", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "reproducibility", "cost_latency_tradeoffs", "governance"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8cdb002cd3066c3ae0d86b9fa6c5b7048b77134c"}
{"id": "train_48361", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "tooling", "documentation"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2084a496545a43f20c292864964e74e69f49af0f"}
{"id": "train_48362", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "auditability", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2be4d25d34abbcc007e124920f92ba9f5fc910ab"}
{"id": "train_48363", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "documentation", "auditability", "security_gates"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "80585e1fae64fbd10d5a35186163d8ebe9af4018"}
{"id": "train_48364", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "auditability", "documentation", "ci_integration"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "187c40531500ce24e220626003d62a774bb1a339"}
{"id": "train_48365", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "documentation", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2e10764236526fee689f15d977307febb6123ae5"}
{"id": "train_48366", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "auditability", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "87551303dcd69e32bc6097730251ac63fdc480fc"}
{"id": "train_48367", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "tooling", "ci_integration", "documentation"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3a654abbc531f0c2bf5f4117d72b38fd5b5bbb66"}
{"id": "train_48368", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "governance", "documentation", "reproducibility"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1a63cdd152c163f5905a8258f3889e176889332a"}
{"id": "train_48369", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "ci_integration", "tests_are_truth", "security_gates"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "10c07a9440493a2db532665880287c523cc9b313"}
{"id": "train_48370", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "tooling", "repo_scale_reasoning", "documentation"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "72742ce671a3521b25afc1a3b396da93775c2525"}
{"id": "train_48371", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "tooling", "ci_integration", "tests_are_truth"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e711eaf4459091ce169e0cfffd8f988dc49f8975"}
{"id": "train_48372", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "repo_scale_reasoning", "auditability", "tooling"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4917d681291af8f8a00f19f7fbcaf617c99176f6"}
{"id": "train_48373", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "evaluation_metrics", "security_gates", "governance"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "63465dcff70f5b810c491b426563577d3dbe0dd3"}
{"id": "train_48374", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "documentation", "governance", "reproducibility"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4b4e4985589d1a4236d581645a18c6a5716488ba"}
{"id": "train_48375", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "governance", "auditability", "reproducibility"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b97088381768bc275c00cb96a0cd196247a84162"}
{"id": "train_48376", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "evaluation_metrics", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8a6e25c68ea92916c7236e5b1d748a2a011d4940"}
{"id": "train_48377", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "tests_are_truth", "tooling", "governance"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "04bbaae5822744ee7dfa0372900683bc2b62c077"}
{"id": "train_48378", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "documentation", "auditability", "tooling"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9330186d3aa86af46b95c560567a63b531c7625d"}
{"id": "train_48379", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "tooling", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "04787b6179262cd4f13abba8b7459a1af1fa3c02"}
{"id": "train_48380", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "tooling", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "979833cfe99d9193ea7e55bef1d15c221f5ecef3"}
{"id": "train_48381", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "evaluation_metrics", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4ac4021dbde693a457d587d845bda02834e9e34d"}
{"id": "train_48382", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "tests_are_truth", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8b33983fb09f205a8387c24be9d26463359405ac"}
{"id": "train_48383", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "tooling", "auditability", "evaluation_metrics"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "20b84484fb4f8de45443f72004017c6bda7e1c26"}
{"id": "train_48384", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "governance", "ci_integration", "tooling"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5fdf321f0b6915c832a275a2a436d58ada10a0c2"}
{"id": "train_48385", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "evaluation_metrics", "repo_scale_reasoning", "tooling"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b97846b75b39dbd6f598adb37865481d27f8982f"}
{"id": "train_48386", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "tests_are_truth", "security_gates", "governance"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bfa3bf7fae0afb6174e7a17d258d2f7e4acd1da7"}
{"id": "train_48387", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "cost_latency_tradeoffs", "tooling", "tests_are_truth"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bbce43d0ba91ed01e1e0d11312daa691ddd5320a"}
{"id": "train_48388", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "documentation", "auditability", "security_gates"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ccbe964c54f0b5722208fc05362631c265a248c6"}
{"id": "train_48389", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "auditability", "ci_integration"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5d05778590b94f6bc360324b68f930383feeb70c"}
{"id": "train_48390", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "security_gates", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "56f52de8eb80d684805e7a273ca0d45b0483f1b0"}
{"id": "train_48391", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "repo_scale_reasoning", "security_gates"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4ab1242e9570313804a4f5fa69d13592ab22bfbe"}
{"id": "train_48392", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "security_gates", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "df7997378c247e89f04b04136740f59f01062413"}
{"id": "train_48393", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "repo_scale_reasoning", "security_gates", "governance"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1f6b99c1e81fb65940d76d693d4be0ce9bf42836"}
{"id": "train_48394", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "governance", "documentation"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7f8f234154349ad407c27e6ff0c93b9569485725"}
{"id": "train_48395", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "tooling", "tests_are_truth", "reproducibility"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9880a9941d0661b112f7ae785d9c5d545eac6b7d"}
{"id": "train_48396", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "evaluation_metrics", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d88454134419e7bdf33789dbc5a9e0c3bdce035c"}
{"id": "train_48397", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "reproducibility", "tests_are_truth"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "26bf17cecd82c3e3d221cbb110688c9f13143f0d"}
{"id": "train_48398", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "evaluation_metrics", "tests_are_truth", "documentation"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d1a44373e98e4d5186b8ea47881c152ed5377b7c"}
{"id": "train_48399", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "security_gates", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "736b9eacfbd5a515b8db4d260d164b37623cd266"}
{"id": "train_48400", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "evaluation_metrics", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "52c8079e530a1246eb4e1405804625110ff991d9"}
{"id": "train_48401", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "auditability", "security_gates", "tests_are_truth"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a28b345e3c712f544ca7d62453c98580642d8f94"}
{"id": "train_48402", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "repo_scale_reasoning", "evaluation_metrics", "reproducibility"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "72cde4bb62c0ecfe135382a7ebf4486b79174688"}
{"id": "train_48403", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ec793e74b7a77238d196132885fd66098eb84bc1"}
{"id": "train_48404", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "security_gates", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2327e1a45f7e6574b065376715fa9bcd4d69c7cd"}
{"id": "train_48405", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "security_gates", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f090c84e592938bb7ca98a8cd324ca7aaed7030b"}
{"id": "train_48406", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "security_gates", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c8556c2808dbed540f43f545ba3751cc772d18c1"}
{"id": "train_48407", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "governance", "reproducibility", "security_gates"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "084fcbfbc11c8e08935b9788e1a80fd792ebaf7c"}
{"id": "train_48408", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "tooling", "security_gates", "reproducibility"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7e360574840239e45834f7ab7dcf54cc2b656a2a"}
{"id": "train_48409", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "repo_scale_reasoning", "security_gates"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6094463eeb47d3c31322131329d51b7930cb6df9"}
{"id": "train_48410", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "ci_integration", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1c3c6e05aade1bcc6fa2c5cc1f647e56c95051fc"}
{"id": "train_48411", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "tests_are_truth", "security_gates", "governance"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9eba037f5528472a0bf10537dd2a71041cb1c204"}
{"id": "train_48412", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "security_gates", "repo_scale_reasoning"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "727c11337c3f496f818eb96dc32e747ce4826a59"}
{"id": "train_48413", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "governance", "security_gates"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "633c8fed64b1d09e5703d5bd7c27d7860d037c65"}
{"id": "train_48414", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "security_gates", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "710b98b67ed87e672c7d5eec1e70c0a22fa447ea"}
{"id": "train_48415", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "reproducibility", "security_gates", "tests_are_truth"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e2d2777dad319b4655be5589d18e17e9ad666df1"}
{"id": "train_48416", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "documentation", "reproducibility", "tooling"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "75903c635f2ff97e3fa1965ed7fa5e2e83f51b37"}
{"id": "train_48417", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "ci_integration", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "66a5dad5e5c759f4beaaec3770e749e643d22370"}
{"id": "train_48418", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "governance", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f8f79f02ec0b62ee002f4ee28e10fc72e1b8fd78"}
{"id": "train_48419", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "reproducibility", "documentation", "ci_integration"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "78ffd0334c2cd9665cb1a39cd38c8a741d7e4be9"}
{"id": "train_48420", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "cost_latency_tradeoffs", "governance", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "85ff536d3f5b2bbcfce3796b57596b5c50684016"}
{"id": "train_48421", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "security_gates", "tooling", "documentation"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fc95aba9767272637685e4a90da6f81ec859df4f"}
{"id": "train_48422", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "repo_scale_reasoning", "reproducibility", "documentation"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ee43ccaac1fbc6f6ccd119aed628638306a37703"}
{"id": "train_48423", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "security_gates", "ci_integration", "reproducibility"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "11cdde58733b63c61f7a72c8674c8d103019fc09"}
{"id": "train_48424", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "cost_latency_tradeoffs", "reproducibility", "governance"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "53ccdf88dd7d15145bf87c5b80de7e1e2915e974"}
{"id": "train_48425", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "tests_are_truth", "security_gates"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "086fb378f84a67d5449906135e62835ae3089dbd"}
{"id": "train_48426", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "cost_latency_tradeoffs", "governance", "reproducibility"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "19eb337c6f85eeb93641e21e5c656f477c70af36"}
{"id": "train_48427", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "auditability", "evaluation_metrics", "documentation"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4d991c6ea4bcd056bc06335b7264f4f87465d1e6"}
{"id": "train_48428", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "auditability", "cost_latency_tradeoffs", "governance"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c2d5804ef4cadc45d5175fa9c0111d93fe01512a"}
{"id": "train_48429", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "repo_scale_reasoning", "documentation", "ci_integration"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "762d7511047cbfffae46949873985dcecfbe13a3"}
{"id": "train_48430", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "evaluation_metrics", "tooling", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "45816a2c8d0fbfa070b537d2b5030fd1be773570"}
{"id": "train_48431", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "documentation", "tests_are_truth", "security_gates"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "711f4c8f6d590ad1b34bf338238de011ee88baa0"}
{"id": "train_48432", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "ci_integration", "tests_are_truth"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "685021de37aa3f6b13dabb59674db273fdbcc8ea"}
{"id": "train_48433", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "tooling", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a0ee28554b6237f11c4a30eadda5691955f58c36"}
{"id": "train_48434", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "reproducibility", "repo_scale_reasoning", "governance"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2c32fa42666b9addc19ca2c456b91adc1ac069a2"}
{"id": "train_48435", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "auditability", "governance"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0a0e89a9ea1abe78f64d2cfc542f921ae8d57bd9"}
{"id": "train_48436", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "documentation", "auditability", "governance"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eb200d83f3a04e4d1986c4ee974a88aeb53e552f"}
{"id": "train_48437", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "tests_are_truth", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3b9495478c6ec53b80cb5ce8e6d92c24e97fda8f"}
{"id": "train_48438", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "reproducibility", "documentation"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4341a5c69b468576b0009345bb1a0427a3799957"}
{"id": "train_48439", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "governance", "tooling", "security_gates"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "95fcce900e805b519c48bba2234fbf98ae49e5e3"}
{"id": "train_48440", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "documentation", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "300d365604bc2b92bf9db096f385cfb7f8beba1e"}
{"id": "train_48441", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "cost_latency_tradeoffs", "documentation", "tests_are_truth"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4571f25fefa3e4567e33f99471f07e952732b211"}
{"id": "train_48442", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "cost_latency_tradeoffs", "documentation", "security_gates"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d1025f47f5057afddaf850ecebd8b101b813b362"}
{"id": "train_48443", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "governance", "tests_are_truth", "tooling"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d22fc1f417c4797cc964c921dc2850bef7515601"}
{"id": "train_48444", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "security_gates", "tooling"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c12b88a227590c48f66ddb4a6a79d98a996f5f78"}
{"id": "train_48445", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "ci_integration", "repo_scale_reasoning", "security_gates"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "95b4acd3469ca3a6799d00f2e5196332d9339a84"}
{"id": "train_48446", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "security_gates", "ci_integration", "governance"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dd1b1c59bfb8e18fc5cb0300fa461d3c936e593f"}
{"id": "train_48447", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["governance", "ci_integration", "security_gates", "reproducibility"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0f2678166d33250fa37f1386a13046b98a0d7faf"}
{"id": "train_48448", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "tests_are_truth", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5a1d6e4f51d08f1d00d40857d3971a736f89b93a"}
{"id": "train_48449", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "reproducibility", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d8baea7632e20691eeb552623a6f929fac133beb"}
{"id": "train_48450", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "ci_integration", "auditability", "governance"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7eec7c8baa3362c38dff58def4651dd1d5cc108a"}
{"id": "train_48451", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "security_gates", "governance", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7cc87ae79683495170804956536e7c5b2752ade5"}
{"id": "train_48452", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "reproducibility", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1c773fc40043744eda4a774d7e43bcc20965bfa6"}
{"id": "train_48453", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "ci_integration", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "16528b8f5dd93d54657e57d168f5d8b01db8f187"}
{"id": "train_48454", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "reproducibility", "governance", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8045f617505289666a97a52789150bd59da2a25b"}
{"id": "train_48455", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "ci_integration", "governance", "auditability"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "645190e36bff4439455a53aab723553401fe0767"}
{"id": "train_48456", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "documentation", "auditability", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8072c8efb4a7f63701d688f2ed8bb08571b016d1"}
{"id": "train_48457", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "ci_integration", "cost_latency_tradeoffs", "governance"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e974df8e1d7ed2513c8c6ddc631b7ff2baae2500"}
{"id": "train_48458", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "auditability", "ci_integration", "documentation"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "49eaeea5ea371744cce084f9600dc1b3f1149389"}
{"id": "train_48459", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "repo_scale_reasoning", "evaluation_metrics", "reproducibility"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1264f1775a4182a8bf87696aa81bd7eeae08b25a"}
{"id": "train_48460", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "documentation", "repo_scale_reasoning", "governance"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "96f959fdc4bfd58652d70a91bec67ebfe5435d0b"}
{"id": "train_48461", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "reproducibility", "evaluation_metrics", "governance"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d7b3b8a90e5e137f9a82e04eb1401842784b09cb"}
{"id": "train_48462", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "ci_integration", "evaluation_metrics", "security_gates"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7012427da08b93070916f45e327d179d47749983"}
{"id": "train_48463", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "tooling", "governance", "auditability"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "99abcc1cac230846fb44439884e10af0c254d604"}
{"id": "train_48464", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "documentation", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "474a061b8e10abdc5be295786429e0c273eed04e"}
{"id": "train_48465", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "auditability", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ced72b525d2ee31b3d8e3c3e1f66182868394830"}
{"id": "train_48466", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "evaluation_metrics", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "297bb6ce10d2b47b155a9974d532e0d5213648c5"}
{"id": "train_48467", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "documentation", "tooling", "governance"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "adf6018b2a441bf89eeb45cc0090ee8c04f01316"}
{"id": "train_48468", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "documentation", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "664eddeb65d064609df0931f53db259e919ddbd2"}
{"id": "train_48469", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "tests_are_truth", "documentation", "reproducibility"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2d44e73e142dffcdf1ff36c0a007ed9a22329fa7"}
{"id": "train_48470", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "ci_integration", "governance", "repo_scale_reasoning"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0ca7db75dadf8dd896bd81f126212c8ca8f81d55"}
{"id": "train_48471", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "auditability", "evaluation_metrics", "governance"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e02079da83184b185d58edb85424f9d4c58611c6"}
{"id": "train_48472", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "documentation", "cost_latency_tradeoffs", "governance"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cb46b44a5bdee05a218ce561055af994be992d53"}
{"id": "train_48473", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "documentation", "governance", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cf4e71d4f16493e21266dd9f46d67d46cb02e6ba"}
{"id": "train_48474", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "governance", "evaluation_metrics", "reproducibility"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "70826dd52247054967a20032a1169505a6019b19"}
{"id": "train_48475", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "tooling", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "889a51edcb34e0656885ea0451eec0f6343c0b42"}
{"id": "train_48476", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "tests_are_truth", "reproducibility", "tooling"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1f6881c253590cf01c3a73a165ef0b56be72f6be"}
{"id": "train_48477", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "security_gates", "governance", "repo_scale_reasoning"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7b025663d8e832fa6e7363f6d1ddd4f7f424e25d"}
{"id": "train_48478", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "documentation", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ecbd121fdcef20b2beafefe311e61be5f1929109"}
{"id": "train_48479", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "ci_integration", "tooling", "tests_are_truth"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "36a66aaacc10516a493f5b2dd9026a52905268c6"}
{"id": "train_48480", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "auditability", "tooling", "ci_integration"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "77be0753357ff2dbff9081fb7fcce13f9363f7a3"}
{"id": "train_48481", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "documentation", "security_gates", "tooling"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5653125fa2d83b724a40407dd5c185b427bf89dd"}
{"id": "train_48482", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "security_gates", "auditability", "governance"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2c869dfc030ec3329bbb9c64afccf46ecbd85f0d"}
{"id": "train_48483", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "security_gates", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "57fa72ad742971faf543830ccf73ed5658171dd6"}
{"id": "train_48484", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "security_gates", "ci_integration"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6b4e8c2688f75a57411ff6590f1075f5ff9f46bb"}
{"id": "train_48485", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "auditability", "reproducibility"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "49c7b2b73ab0073fe8c9e6d046a166a2bdf00d60"}
{"id": "train_48486", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "governance", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6343f9e397f886063fc03c55c5cc48ceff496cbd"}
{"id": "train_48487", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "reproducibility", "tooling", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8dcf8c926bc8d62b48140cfc47c131c2f6083253"}
{"id": "train_48488", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "governance", "tooling", "evaluation_metrics"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4876c7627bec5804faa8c790dc1ac1ad5dcaf4a3"}
{"id": "train_48489", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "auditability", "reproducibility", "tooling"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2fb6b1be5ea9c35e2f48932785a577584c98cae8"}
{"id": "train_48490", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "auditability", "tests_are_truth", "governance"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6fd25e0ac97842f1b52b47bcb68219ce3e6dc57e"}
{"id": "train_48491", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "evaluation_metrics", "ci_integration", "tooling"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a655e24d71cadb2f5e3f279f869b836f9b0e537a"}
{"id": "train_48492", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "documentation", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d03d62cc1bf48ffbe99d0d0e3e5027575d3f6e45"}
{"id": "train_48493", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "security_gates", "documentation", "tests_are_truth"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "49fef118e5ca2ec579743bf54f10c699ed9d7b7a"}
{"id": "train_48494", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "security_gates", "tooling", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3a53293a4d826d62e4a560d1ea66c7d57b76498b"}
{"id": "train_48495", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "tooling", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "165c815b06bfa31ddc74ae029a0627ea5b3a324d"}
{"id": "train_48496", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "ci_integration", "documentation", "auditability"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0e9300ef846366a9963e85dbeed23c2d33d40899"}
{"id": "train_48497", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "tooling", "governance", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "885b73de96d5edf4f722d74633b6546426ebadfd"}
{"id": "train_48498", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "security_gates", "evaluation_metrics", "ci_integration"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "20a05f08aaa8fc44d65d041918d65c2063ed57ca"}
{"id": "train_48499", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "reproducibility", "auditability", "documentation"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ade457347bf1f1f5a7994c5523cbcf29b5c32b85"}
{"id": "train_48500", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "ci_integration", "auditability", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d80734768b710b8f78d8f250782f9efd351f96db"}
{"id": "train_48501", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "auditability", "tooling"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "23e339235150089ca56e698cab96a597f093dbbb"}
{"id": "train_48502", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "cost_latency_tradeoffs", "documentation", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0e0983256e69717b5018610d52eaf4b71f2f960d"}
{"id": "train_48503", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "reproducibility", "auditability", "tooling"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "98aaf9e0feb8b0330d4723bea39cb6ae4ec3f867"}
{"id": "train_48504", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "reproducibility", "evaluation_metrics", "governance"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f1930328c2742b284038513464a7ea111d7e464c"}
{"id": "train_48505", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "auditability", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d6bfdf3f2166ad012f89211d5a7cf232187b45ff"}
{"id": "train_48506", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "tests_are_truth", "documentation", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1aceab1a02331f71eec01114eb3e206b78f45e0d"}
{"id": "train_48507", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "tests_are_truth", "reproducibility", "documentation"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a72e29965e59aadd1b4a4eacb6d5156ab0db3b57"}
{"id": "train_48508", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "ci_integration", "documentation"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f86f21ff3c4dadcce1e18c09f7e11d7329474f92"}
{"id": "train_48509", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "governance", "reproducibility", "security_gates"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f27fdf6511d049cbc62aa177a76974b5df655686"}
{"id": "train_48510", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "tests_are_truth", "reproducibility", "security_gates"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "794190f36cb689e607aaa2977c577a250ec55c41"}
{"id": "train_48511", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "ci_integration", "security_gates", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7b1676a0a99ced1d138c650f52d071298d88e9e2"}
{"id": "train_48512", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "ci_integration", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "895b1e1cc947347695d8e694ba91181d79630187"}
{"id": "train_48513", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["governance", "evaluation_metrics", "reproducibility", "tests_are_truth"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b84a515a0ab83576db43463451ae835805af2478"}
{"id": "train_48514", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "ci_integration", "auditability", "tests_are_truth"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8a1610bb85516d82c5df8fb0bda4d47b1abbb720"}
{"id": "train_48515", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "reproducibility", "security_gates", "tooling"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e69dd797687dd8726ae30bbea84094697b7bdfd4"}
{"id": "train_48516", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "evaluation_metrics", "tooling"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a5859e6a49df1135e92573ab00d3e480f5cec1a8"}
{"id": "train_48517", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "auditability", "ci_integration"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "086fb378f84a67d5449906135e62835ae3089dbd"}
{"id": "train_48518", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "tooling", "repo_scale_reasoning", "security_gates"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dcf9fd7cc2a550e5e3f0080899c8ab2076920ddb"}
{"id": "train_48519", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "cost_latency_tradeoffs", "documentation", "ci_integration"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9af398ceb135ec07e106d90d4362a500f36c23a3"}
{"id": "train_48520", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "documentation", "tooling"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "95e6cf5a69b45016cf703ee9145ee6644f9b6b95"}
{"id": "train_48521", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "tests_are_truth", "security_gates"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0eabf520cea127a29430e5f4e385f2c8d6b4f193"}
{"id": "train_48522", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "ci_integration", "auditability", "tests_are_truth"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5d03ebc4f2f209516718874f0f02506d5cb2baaa"}
{"id": "train_48523", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "auditability", "evaluation_metrics", "tooling"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1d8fb92b815e2d3c8957565b3ad01c8c3b106f52"}
{"id": "train_48524", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "tests_are_truth", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0a900d217452a73ba952efcfaa678cd0a7f04b09"}
{"id": "train_48525", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "auditability", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4880967ec84cf26cd31fbed45fd38bd58471b818"}
{"id": "train_48526", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "reproducibility", "security_gates", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d006ac430b5e1d79f0f92222ea12ea077846f78e"}
{"id": "train_48527", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "ci_integration", "security_gates", "tooling"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5993ad71f417e23d6dd64453be4236defd0de693"}
{"id": "train_48528", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "governance", "auditability", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0867cf4476cc7cb09528794770dc800665dbf536"}
{"id": "train_48529", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "documentation", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9470f88bbf7655cda691db766524009f00cbdb5e"}
{"id": "train_48530", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "governance", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ffcbe80116ec4ba122b71b76eb0e99e47cd58f46"}
{"id": "train_48531", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "evaluation_metrics", "tests_are_truth", "governance"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "268e41bc8fbb5397824dc23c175066f15cc2b98c"}
{"id": "train_48532", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "tooling", "ci_integration", "reproducibility"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7531a5a7be2bd2420e163d2853eb77df1c1cfa28"}
{"id": "train_48533", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "documentation", "repo_scale_reasoning"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "50933541e70ebf4446d31d20efc2c0533ebca47e"}
{"id": "train_48534", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "evaluation_metrics", "reproducibility", "governance"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9fdfbb0ba4fdea10a7471bb2b1c93979bea640f3"}
{"id": "train_48535", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "governance", "documentation", "tests_are_truth"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c8f5401e9d327aac096856adbe3ab1918995fa71"}
{"id": "train_48536", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "governance", "tooling"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "691e3f51693564664f1dce49db364981383e48bf"}
{"id": "train_48537", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "cost_latency_tradeoffs", "documentation", "governance"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ab7800b3fc708cc87011c82f2621d707a822159f"}
{"id": "train_48538", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "auditability", "repo_scale_reasoning", "security_gates"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d64c37c649623f50ab92a59fb5c64b6549cc7ddc"}
{"id": "train_48539", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "reproducibility", "documentation", "tests_are_truth"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e498ae8b098874941974cff60ae1b3fff94bd02e"}
{"id": "train_48540", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "reproducibility", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e349b3de38c77b62dc52c3890cb5a2e55a3b4232"}
{"id": "train_48541", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "reproducibility", "cost_latency_tradeoffs", "governance"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "af5c46e41f846a80ca8ec2fb14cece0a0c9d956a"}
{"id": "train_48542", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "reproducibility", "documentation", "auditability"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f4643c2174fe239473eebab5362cf2291c6af1d1"}
{"id": "train_48543", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "documentation", "repo_scale_reasoning", "governance"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "72bf759fb93ac00db06437933147be5f874353ff"}
{"id": "train_48544", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "security_gates", "reproducibility", "tooling"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1a0960e7e240d93684fcb891349bb29fefba49fb"}
{"id": "train_48545", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "tests_are_truth", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ccd800a33f5b06ac7078cdd6187b5ba3b4d82f19"}
{"id": "train_48546", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "documentation", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fbe5e522145856f09d13c667265df40a7c40026c"}
{"id": "train_48547", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "evaluation_metrics", "security_gates", "ci_integration"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a4dc146e42fc87d8a6a4c8263673bdf9967422f1"}
{"id": "train_48548", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "ci_integration", "reproducibility"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "43d8e8c11339bc0a38b05cf4ca7cc05ea105fd60"}
{"id": "train_48549", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "governance", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "48ba12da3d2855f46ed5477f4c87cded3af6864b"}
{"id": "train_48550", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "reproducibility", "repo_scale_reasoning", "auditability"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "882f83b546a7264dafd9b26949c5ff6070f072ec"}
{"id": "train_48551", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "repo_scale_reasoning", "governance", "tests_are_truth"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0fe6b44f28de1008e8b3cda7a7b3595d225f81bd"}
{"id": "train_48552", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "reproducibility", "governance", "ci_integration"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c477bb993d55082cb8605ae3ae2d4dcb5f2ae692"}
{"id": "train_48553", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "tests_are_truth", "security_gates", "governance"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b39916ea39e6e511f2043ee476059da4ede82d85"}
{"id": "train_48554", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "governance", "documentation", "reproducibility"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f2dc3d8eb12bde8a684ae810b266a994080eb57f"}
{"id": "train_48555", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "tests_are_truth", "governance"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b812686e8c27e8450f856d4f93788b1f1a5ccf2e"}
{"id": "train_48556", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "evaluation_metrics", "auditability", "security_gates"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0998f0630ab08b6d3a0957d4c22d7ef34e3492d0"}
{"id": "train_48557", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "reproducibility", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ccc727e7523625114a2b47edb7cfaa4e8006b64b"}
{"id": "train_48558", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "evaluation_metrics", "repo_scale_reasoning", "security_gates"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "76f058a0eb8512f6c6a10a5b2392814cbd4a4402"}
{"id": "train_48559", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "cost_latency_tradeoffs", "governance", "ci_integration"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "88f567a1fea656883b89d1e2690ba1b99d618b35"}
{"id": "train_48560", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "documentation", "tooling", "governance"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6cdd52a9c373cc0374621f52e8329c7e27d607c6"}
{"id": "train_48561", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "81de0ed2861a9618363dbb7041ceee69cb7c83cb"}
{"id": "train_48562", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "documentation", "evaluation_metrics", "reproducibility"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "78249f821dcffa92a3d1ff9138a5398a5367f08c"}
{"id": "train_48563", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "tests_are_truth", "auditability", "reproducibility"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f536e21a4198efc3edad42b678f48be4e4cdaace"}
{"id": "train_48564", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "governance", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "992f9d13724fc1c7761538c825d81e3dc3afeb38"}
{"id": "train_48565", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "governance", "documentation", "tests_are_truth"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3fb429210a1efb0b98d0b04fde4c655f18425085"}
{"id": "train_48566", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "governance", "ci_integration", "documentation"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e96beddad5f2ee2870bb09b14812d976de30467a"}
{"id": "train_48567", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "repo_scale_reasoning", "security_gates", "tooling"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dc52838f40ac6aaf5c4db9ed8b727aab6031d074"}
{"id": "train_48568", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "governance", "security_gates", "tests_are_truth"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eaa1b8ec4e5fa44870560ca5bc76674bea0703f6"}
{"id": "train_48569", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "documentation", "governance", "security_gates"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4b82abec5e7ec40c919e566f4c00df056ac54b91"}
{"id": "train_48570", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "repo_scale_reasoning", "ci_integration", "documentation"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "702853629ad2459f546e6d14cd212148d1169f54"}
{"id": "train_48571", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "evaluation_metrics", "reproducibility"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "341720eefcd0cccf33261849801517458e9f0d13"}
{"id": "train_48572", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "governance", "tests_are_truth", "auditability"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "61a9e45eb876cfe83f4b9ad6b3a4904346fb43cc"}
{"id": "train_48573", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "repo_scale_reasoning", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "56cdb46f3db561f744600a366393232912c86c69"}
{"id": "train_48574", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "documentation", "reproducibility", "evaluation_metrics"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "44f320c89e0301bc14b57a17d81df08fb04231f7"}
{"id": "train_48575", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["auditability", "evaluation_metrics", "ci_integration", "tooling"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8b9324d966f10c01f689a810bfae40711820c527"}
{"id": "train_48576", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "repo_scale_reasoning", "auditability", "ci_integration"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2c99db5669b1592b572f35ee83591b1d31e1313e"}
{"id": "train_48577", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "ci_integration", "cost_latency_tradeoffs", "governance"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "735db2e5f6c2c3163491b5bc41dc3b7e0b91e178"}
{"id": "train_48578", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "cost_latency_tradeoffs", "reproducibility", "auditability"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bbb07585b418e1994663bb161880866bf0e0265f"}
{"id": "train_48579", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "ci_integration", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "14b4fe3773be8d0a6f0d989a19a4aed6ff648840"}
{"id": "train_48580", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "tests_are_truth", "reproducibility", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "68a694b82aab70114fefe3fd2bbf1cebed6ea09a"}
{"id": "train_48581", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "tests_are_truth", "security_gates", "governance"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4ace36ef8bd6a650cefb315fe4b2318e051f1bbf"}
{"id": "train_48582", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "governance", "tooling", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ee8b6c62166ea06da29f6551e19a61fb3150c3ea"}
{"id": "train_48583", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "auditability", "documentation", "tests_are_truth"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3efaab7ae895c58b6a1bdb0a47da1fa5001ed48c"}
{"id": "train_48584", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "auditability", "governance", "documentation"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4f6fc01269f0bc9bcccc117274799399143b9de2"}
{"id": "train_48585", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "evaluation_metrics", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1b6eb74ba605a482125a046f7486788cef4cb2e3"}
{"id": "train_48586", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "tests_are_truth", "documentation", "security_gates"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d539cb6b503702c1b40572c1adc0cc435321d46a"}
{"id": "train_48587", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "ci_integration", "tooling"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "be10c36d5db9b6ab448e22e136c4e1e186cc71eb"}
{"id": "train_48588", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "evaluation_metrics", "cost_latency_tradeoffs", "auditability"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b2642df3d327f2748eab3576d8185adc5821d541"}
{"id": "train_48589", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "auditability", "reproducibility", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9687604bafe64fa885857b0e8253766b43d76b42"}
{"id": "train_48590", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "governance", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eab87991636e04572dbe4fddbc6c6e10c409ccb5"}
{"id": "train_48591", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "tooling", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fe1a214f61e5bd01a9e6835866337ca00c8e499c"}
{"id": "train_48592", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "auditability", "tooling", "tests_are_truth"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bff8235a920423041ae75eef44bb46967496794e"}
{"id": "train_48593", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "tests_are_truth", "auditability", "documentation"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f7542f51d3ce8145be64b5a8836c12cb4b4d8d2a"}
{"id": "train_48594", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "auditability", "documentation", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ca7011b050cb9f87521f050b3cbf6134c79eb1a3"}
{"id": "train_48595", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "auditability", "ci_integration", "security_gates"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3d71c6f8311872cf37e5f65f105ec798a0ef36b4"}
{"id": "train_48596", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "tooling", "auditability"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e0509ed9c7a40ea181c881adf2cc3bb69ffe20cc"}
{"id": "train_48597", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "repo_scale_reasoning", "tooling", "governance"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9a4a5048ca10e7709d6df5853b33ca894d2130cc"}
{"id": "train_48598", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "repo_scale_reasoning", "tooling", "ci_integration"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "79e1465888fc460128f6f909a03cb68c2b19fcd8"}
{"id": "train_48599", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "tooling", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "79a2a8bda7176b57bc53ee1569305f821c66cbdc"}
{"id": "train_48600", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "governance", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "17938f1fc9a85a7698054e032b14593674810d6a"}
{"id": "train_48601", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "governance", "evaluation_metrics", "reproducibility"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "974eb9bb0e474efa52d7ba6bd689913fe242015b"}
{"id": "train_48602", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "tests_are_truth", "auditability", "governance"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d4e5192e8d9ff011858a6b8940017f03f6860384"}
{"id": "train_48603", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "auditability", "governance", "tooling"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "df1e91a19b9e772794eb5c23451a82792260d12f"}
{"id": "train_48604", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "auditability", "ci_integration", "security_gates"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aea77666479a011b49902461644f6f1b0356df2f"}
{"id": "train_48605", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "tests_are_truth", "documentation", "ci_integration"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2deddfc88325b4a15fc2888a80192d246104b6df"}
{"id": "train_48606", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "reproducibility", "governance", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "31119fc39547d5d1ffe9b50fecd13dad057b9352"}
{"id": "train_48607", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "security_gates", "ci_integration", "documentation"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5d6ddcb3bf19a13f8cd102fee0d2c7e6e60d3847"}
{"id": "train_48608", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "tests_are_truth", "auditability", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "80a3f1cac5cd068508d0437ca63393d07913444c"}
{"id": "train_48609", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "evaluation_metrics", "security_gates", "documentation"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f00a85d5180da91ccc97d7cab8b53f15987acb80"}
{"id": "train_48610", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "tests_are_truth", "ci_integration", "reproducibility"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ae9949ad5c54a1749f901b36af1919218ecf971f"}
{"id": "train_48611", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "documentation", "ci_integration", "tests_are_truth"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "162cb4c07ae768bfeb689e53cbc98b616f0068bd"}
{"id": "train_48612", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "evaluation_metrics", "auditability", "tooling"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c5ae34e291fff45b01956667f871bf678d6e1f2d"}
{"id": "train_48613", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "ci_integration", "documentation", "governance"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fb0c6b9988fccb9365b3dfecc71c7bb857043e8e"}
{"id": "train_48614", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "documentation", "auditability", "governance"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "29e1f95b66e30ea631dd783c87428f670252471f"}
{"id": "train_48615", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "governance", "tooling", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e519c5f9aa05b5dd8ac993d2505d86473b1624a9"}
{"id": "train_48616", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "reproducibility", "security_gates", "tooling"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4e61f34d051a820f86983c5b1cfe8d5a489b9947"}
{"id": "train_48617", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["reproducibility", "auditability", "documentation", "tests_are_truth"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d21071107ece7c2410e8d5eef9ae129b4f42d7e5"}
{"id": "train_48618", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "governance", "tests_are_truth", "auditability"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e68a0da6665682b104a06fdfb380eaea0dcfa6da"}
{"id": "train_48619", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "ci_integration", "governance"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "08075d1769c9b7f650509805a744455b38092779"}
{"id": "train_48620", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "ci_integration", "tooling", "tests_are_truth"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "71d4d1c78e012139534f211bcbd50c9f0357b6d0"}
{"id": "train_48621", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["governance", "tests_are_truth", "auditability", "ci_integration"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc2caf6ee9a23d0b37dab141d8d473614432756a"}
{"id": "train_48622", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "tests_are_truth", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e6c145cef6418ee8f71c83374951789b20d8f380"}
{"id": "train_48623", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "tooling", "tests_are_truth", "reproducibility"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b993f714a59d146fd79940306b6dec267a22ecfe"}
{"id": "train_48624", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "documentation", "security_gates", "governance"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "38d1a2b080418d6b824199cf8a58cae0b950d97e"}
{"id": "train_48625", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "ci_integration", "governance", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "61d94a95a73d35a84ce1feb23c5a56939b3b6449"}
{"id": "train_48626", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "29d6d65be21987cd963afed9b3690dad38090927"}
{"id": "train_48627", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "security_gates", "auditability", "tests_are_truth"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "12d10c0ca91dd1e073e18d87446eaf553c9b1387"}
{"id": "train_48628", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "tests_are_truth", "documentation"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0ec4fb6ffbd5fc303f8996bff3ec87093ea66a4f"}
{"id": "train_48629", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "evaluation_metrics", "documentation", "governance"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8573db12577bc47d449d78a0044464068155ca5b"}
{"id": "train_48630", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "governance", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4e6317c0dd0edd95e23cc51fe226632da848f332"}
{"id": "train_48631", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "tooling", "security_gates", "tests_are_truth"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f63f270b2218f2a8760c10d9f35b04511ee1c016"}
{"id": "train_48632", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "auditability", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4db337c377d8ca6ee278b18fec98ea1bae97c79d"}
{"id": "train_48633", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "repo_scale_reasoning", "security_gates", "ci_integration"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c1099d451e8958a3fbdffe50de9c7088a0df86c0"}
{"id": "train_48634", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "tooling", "reproducibility", "documentation"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c5cc9a7a29ce447a31b9d6976d13dacbd41281f4"}
{"id": "train_48635", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "security_gates", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "728970725c8e28ad64b3289a8fbea63d26c66b51"}
{"id": "train_48636", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "repo_scale_reasoning", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "887a8eff45f926ba2620c5d1f350ea779a47d588"}
{"id": "train_48637", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "reproducibility", "tooling", "ci_integration"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9afa47e8f9962338525d51555d2e45659a124bb8"}
{"id": "train_48638", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "ci_integration", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "87ffd06b18fe05d88f777f68b9dafb65659945cc"}
{"id": "train_48639", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "ci_integration", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e8bac355c26d828661d3872576419df977448ed3"}
{"id": "train_48640", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "documentation", "evaluation_metrics", "reproducibility"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "460df273a8cd7744bc56dd28c8237deb7106d7ff"}
{"id": "train_48641", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "governance", "security_gates", "tooling"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b18f1c4d212e72ca5b82b2aff030fb28e3dddefa"}
{"id": "train_48642", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "evaluation_metrics", "tooling"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f2226a5f62aeae03f32d569619a539033f30716c"}
{"id": "train_48643", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "evaluation_metrics", "security_gates", "reproducibility"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0cfc5b16a2b22dc032562cfcf80a9cd75880ee5d"}
{"id": "train_48644", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "ci_integration", "cost_latency_tradeoffs", "governance"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "201bf1e41c457917c6d078b25e1c41185e5b4fef"}
{"id": "train_48645", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "governance", "auditability", "evaluation_metrics"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dbfb42d14d843819b5bff66f9f35a69cfaac56c7"}
{"id": "train_48646", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tooling", "evaluation_metrics", "auditability", "governance"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d078c570307305a501746be2d5f05463b43d2aaa"}
{"id": "train_48647", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "auditability", "repo_scale_reasoning", "governance"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c740d5d37df90310bd881cf67a330bf3691d9bcd"}
{"id": "train_48648", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "cost_latency_tradeoffs", "governance", "reproducibility"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c85b1cacf2d6ca59e52c48fa89646dce1faf0987"}
{"id": "train_48649", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "tests_are_truth", "evaluation_metrics", "governance"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a4b37c104771146f36aa9b7db5a2393cd733728e"}
{"id": "train_48650", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "security_gates", "documentation"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "571e2f640e0c12414f1070dce0a89dbafcd6380d"}
{"id": "train_48651", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "tooling", "governance", "reproducibility"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ea9c267af746c0515192a86ebf69157824a3c260"}
{"id": "train_48652", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "tests_are_truth", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "55afc3f60d169d2af6600f468b9aecd6ae936787"}
{"id": "train_48653", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "reproducibility", "ci_integration", "tests_are_truth"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f3beb7d557b342109b67cfaac5834c28e98a4c28"}
{"id": "train_48654", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "auditability", "reproducibility", "tooling"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ea11daaec00571dd191b656a1b4555ce1782d2f8"}
{"id": "train_48655", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "documentation", "governance", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7f86a160aa999d9aa5554e0cddce18f1bab9c902"}
{"id": "train_48656", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["repo_scale_reasoning", "ci_integration", "tooling", "security_gates"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1dae1e145a3c3ad6f65cdaf50ce1efcbb60371ca"}
{"id": "train_48657", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "documentation", "security_gates", "reproducibility"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bb9ae047fba5f4c65a77f5f0591d27668eed1525"}
{"id": "train_48658", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "security_gates", "reproducibility", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f40db5a87779abe48be3c4a4676c91482da32f90"}
{"id": "train_48659", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "documentation", "reproducibility", "tests_are_truth"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fbebb0f32a5ec09cc90ad31cfe0415062c7a1f76"}
{"id": "train_48660", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["governance", "documentation", "reproducibility", "evaluation_metrics"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "300d365604bc2b92bf9db096f385cfb7f8beba1e"}
{"id": "train_48661", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "auditability", "tests_are_truth", "ci_integration"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "179bc2a19c75e8a36b293f43c0b40a4d4d7a6364"}
{"id": "train_48662", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "cost_latency_tradeoffs", "evaluation_metrics", "reproducibility"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "73b57528beabe45cd7458a97232a9c5c403d5044"}
{"id": "train_48663", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "governance", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "44edcca5bdfe6dc0e2f4f9692f89ca58e11764f9"}
{"id": "train_48664", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "documentation", "security_gates"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0e02bcd0998270608e4e7f74883ff4e7d5863f17"}
{"id": "train_48665", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "ci_integration", "documentation", "tests_are_truth"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "20f76e664701400ea017c042cabeb9e70fe979fa"}
{"id": "train_48666", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "reproducibility", "tests_are_truth", "ci_integration"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c2f35b65257e47501dab669d0b2872a0d937d84e"}
{"id": "train_48667", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "security_gates", "ci_integration", "documentation"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b316869ba3d565b157cfaaf088d05d044603ded2"}
{"id": "train_48668", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "governance", "ci_integration", "evaluation_metrics"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f1b4664aeae25e2211874d7252b4a000f5a8eca5"}
{"id": "train_48669", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "governance", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f5d3f16df9725bed39ae5de1828dfe68e345bc4d"}
{"id": "train_48670", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "documentation", "repo_scale_reasoning", "auditability"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "16efe51db83c70b9a821edb323ffbf459fc1e828"}
{"id": "train_48671", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "auditability", "documentation", "reproducibility"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0da6a45a49ec708bc8ea3ec05686727bec0b6e41"}
{"id": "train_48672", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "tooling", "evaluation_metrics", "ci_integration"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6cf561bd315d1d7154574b4f36bf0a5cf2cee70c"}
{"id": "train_48673", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "tooling", "documentation", "evaluation_metrics"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "92124498431062885fab14b148061ec6105b3677"}
{"id": "train_48674", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "security_gates", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2ed3d244c5e91f249137ecb5c3522f2bdd86f9cf"}
{"id": "train_48675", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "cost_latency_tradeoffs", "reproducibility", "governance"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "65b60ebd808ca65d492450c645e77d0b0946a080"}
{"id": "train_48676", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "tests_are_truth", "cost_latency_tradeoffs", "governance"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0fc249e1da72bec3cd6dde33f6856b19e0d9ba17"}
{"id": "train_48677", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "security_gates", "tooling", "evaluation_metrics"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1155cf217b74f8a0a64b9561d763879209265c2f"}
{"id": "train_48678", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "tooling", "ci_integration"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "31b2bae43143e686aea1c2ca0409d3723e103b87"}
{"id": "train_48679", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "ci_integration", "security_gates", "reproducibility"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "240a65b05452b964c4566014741d5297f8b2b750"}
{"id": "train_48680", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "cost_latency_tradeoffs", "auditability", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a28fb994bc7bf267ce826c3b22cb5c5efc2cad97"}
{"id": "train_48681", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "ci_integration", "evaluation_metrics"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "84113ec124250e287a4bf4405be0a20ffc9644f7"}
{"id": "train_48682", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "ci_integration", "tests_are_truth"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4b3d8ec035ac590bdcd83876a36b251701cf340f"}
{"id": "train_48683", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "tooling", "security_gates", "reproducibility"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "94c31ed0bec25b1761ac0d6a4e161364c789e814"}
{"id": "train_48684", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "security_gates", "evaluation_metrics", "auditability"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "637afcb0e3abbd25f84edba8c52e7007083a1f06"}
{"id": "train_48685", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "tests_are_truth", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a06f9876030402d86d79e1cff4820b19618bba97"}
{"id": "train_48686", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "repo_scale_reasoning", "evaluation_metrics", "governance"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5861c44ae0af941b68f3e713b5ea7576a7a60210"}
{"id": "train_48687", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "ci_integration", "auditability", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fc28c0bd073533976a336f78be52cfee7c07c2d0"}
{"id": "train_48688", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "repo_scale_reasoning", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "73092c1041d3bd6f6b31fb79aacf0186515b2ab2"}
{"id": "train_48689", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["reproducibility", "evaluation_metrics", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f7b0d99b541592fce004e4bc2cfee5a93d7f12e4"}
{"id": "train_48690", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "documentation", "evaluation_metrics", "reproducibility"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b0170cb245c2e86099f7a3757bc141f4ac3a8ea4"}
{"id": "train_48691", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "documentation", "tests_are_truth", "security_gates"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9c94119a66b090013753f8bd39bbb929c8c075b8"}
{"id": "train_48692", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "ci_integration", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f1930328c2742b284038513464a7ea111d7e464c"}
{"id": "train_48693", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "repo_scale_reasoning", "auditability", "reproducibility"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9a868009b99c4c8fcce208aa4c32d81c86b3df7e"}
{"id": "train_48694", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "cost_latency_tradeoffs", "reproducibility", "tooling"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c89ef98053ca1870638f0c2286a3a08a373d9799"}
{"id": "train_48695", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "documentation", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f30915c15bebb00882c362b69e16e0d36d493fdb"}
{"id": "train_48696", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "security_gates", "reproducibility", "tests_are_truth"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d39782139042eab69609be5158aa007297725857"}
{"id": "train_48697", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "security_gates", "auditability", "tooling"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1f7045bf682ac24f190b147e052e1e9542e7e0da"}
{"id": "train_48698", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "governance", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8258a50b8f1dae2ce1c77c38aad0305145c2f9fc"}
{"id": "train_48699", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "reproducibility", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7a440e929f8b37bd7ece14360ddcf1a2a733065d"}
{"id": "train_48700", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "tooling", "auditability", "security_gates"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eb568292982c8f54b8b7f25f82fb9e0b651452a6"}
{"id": "train_48701", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "tests_are_truth", "security_gates", "ci_integration"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "99abcc1cac230846fb44439884e10af0c254d604"}
{"id": "train_48702", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["ci_integration", "reproducibility", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "74950df59114b75a8273ddf9b1686cf9e1316486"}
{"id": "train_48703", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "documentation", "tooling"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3559fe30e936d964e942119ad751a86d4f88a46a"}
{"id": "train_48704", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "documentation", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2b007d3f744ed4c83f0f8fba87b734e27981444f"}
{"id": "train_48705", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "repo_scale_reasoning", "ci_integration", "security_gates"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "36acd1af1e977f509eff45084cafccaf2dd46961"}
{"id": "train_48706", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "evaluation_metrics", "documentation", "tooling"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e359925ccd56861a3987a222e6b4bec1e4dc37d4"}
{"id": "train_48707", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "documentation", "security_gates", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "81960c30143819dbd172fd8121980f4bd4d08912"}
{"id": "train_48708", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "ci_integration", "reproducibility", "security_gates"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "820d6353605c686c9a0ec8a7d8644445c2442c1f"}
{"id": "train_48709", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "security_gates", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bdc787d44fa0e3bf5bd516439d181a2de266f37e"}
{"id": "train_48710", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "ci_integration", "governance", "tests_are_truth"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4eec22863b105b759a5dd0e80752b4d609c8a825"}
{"id": "train_48711", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "ci_integration", "tooling", "documentation"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0923080ed5073f3f5e2692a4c08cdda61585390b"}
{"id": "train_48712", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "tests_are_truth", "ci_integration", "auditability"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "414a34330b5681211c388e3beeaf738a2a58b152"}
{"id": "train_48713", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "reproducibility", "tooling", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3ea4cb1f05cb9eed72e8f6c49b94be1a747dcb27"}
{"id": "train_48714", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "tests_are_truth", "auditability", "ci_integration"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "53b701217f19b23a1e796955155004ec1aa63b2e"}
{"id": "train_48715", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "repo_scale_reasoning", "evaluation_metrics", "ci_integration"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7271764c7d592edeea7fbfe13931b44c934e11be"}
{"id": "train_48716", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "repo_scale_reasoning", "auditability", "ci_integration"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "26616fd2da03a76fd2cc0370a68730cc895488e6"}
{"id": "train_48717", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "security_gates", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a574c415142b061c08024e6738167c853cf5c327"}
{"id": "train_48718", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "tests_are_truth", "security_gates", "tooling"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8ebc897824efa75147612af6472b0b63e0bfefdb"}
{"id": "train_48719", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "auditability", "governance", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "701bfcf700bd40e0a825f5d77feb151b5dd7b6ba"}
{"id": "train_48720", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "reproducibility", "governance"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3cca022a649cfacd5c5da169269a70faf7ca1cad"}
{"id": "train_48721", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "reproducibility", "auditability", "repo_scale_reasoning"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0a5aa840de96dbece231bf84ea462b7e67a6250f"}
{"id": "train_48722", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "auditability", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ed2f30ea1bda92a655ba2842575df87a7ee19633"}
{"id": "train_48723", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "documentation", "evaluation_metrics", "auditability"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3efcdce113bea25ac7722db58c1eb0314a74cbb2"}
{"id": "train_48724", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "tooling", "ci_integration", "documentation"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "167766f7abbdc021b1502417abe33aee105ab80c"}
{"id": "train_48725", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "auditability", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "987516d6c148eb6acd120e005f5f6b9f2615f609"}
{"id": "train_48726", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["security_gates", "reproducibility", "auditability", "ci_integration"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bea4c214cf1a535cf8e3bfb63d9638eeb8a485dd"}
{"id": "train_48727", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["reproducibility", "documentation", "tooling", "evaluation_metrics"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "81f67d5f3ce3277ff1233f341d7b9aae94d37ba4"}
{"id": "train_48728", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "documentation", "auditability"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f368addcb0999a645dccfd18d00cc429f96b2a41"}
{"id": "train_48729", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "governance", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f34eca0edb6e097298c19057d0ea74d54a8bd0d5"}
{"id": "train_48730", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "documentation", "ci_integration", "evaluation_metrics"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0289e9f97c1e9e5f98c6b90c76401acfc51d9471"}
{"id": "train_48731", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "governance", "documentation"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "89ef97748a4eac217bd3d775612edabc4cc04c05"}
{"id": "train_48732", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "governance", "tests_are_truth", "tooling"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5adc8701953ceb7831b7334e9dcc0dc1a69a94da"}
{"id": "train_48733", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "tooling", "governance", "ci_integration"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1a9667fe81df646d88cb52f23b96f783fab464bc"}
{"id": "train_48734", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "auditability", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a3ee21fc98941bcf7f2d286b070d73da8c58d8ef"}
{"id": "train_48735", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "governance", "auditability", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5f9b927e0bcf7a8a815ae327449ae36d2812ef58"}
{"id": "train_48736", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "tests_are_truth", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "191f89d51ddafa2fc97d018445ba00133b48aca0"}
{"id": "train_48737", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "auditability", "ci_integration"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7415773381e5eeb1e3d353f4dbf4f5041dc31454"}
{"id": "train_48738", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "documentation", "evaluation_metrics", "auditability"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b32617d58afe2471ba508c2c4c42cbc7ae84521d"}
{"id": "train_48739", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["ci_integration", "evaluation_metrics", "security_gates", "auditability"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c2acdb6593480244b13f0c63e8a04bd40b4ecf33"}
{"id": "train_48740", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "auditability", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4a3e25d9fe0388745d2ae56e2f07d7dc46017749"}
{"id": "train_48741", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "security_gates", "cost_latency_tradeoffs", "governance"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "60f65e90b7990a89017dd7233e5d636e25255ec8"}
{"id": "train_48742", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "evaluation_metrics", "ci_integration"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "969b47d187e87cf0e2e1786d158ec4034523d8b2"}
{"id": "train_48743", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "auditability", "tests_are_truth", "tooling"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "86cad0b0cfb8f71735785306ceaf48394e3e2ed5"}
{"id": "train_48744", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "governance", "tests_are_truth", "reproducibility"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "af813ccd590c2c4b5a176fbde85765c30cd02593"}
{"id": "train_48745", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "tests_are_truth", "governance", "reproducibility"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a51d2c4ad264325b307dfa797e3f5ade7d8616fd"}
{"id": "train_48746", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "evaluation_metrics", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6ec24ca38c29e926e548dc409d6d1833e4c326a1"}
{"id": "train_48747", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "tooling", "ci_integration", "reproducibility"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7fcb9771a9d0ec056b5f0b723d5818e4d391123a"}
{"id": "train_48748", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "ci_integration", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bf67d958c3ea026f9610a75953fc476ee8eca17a"}
{"id": "train_48749", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "tooling", "reproducibility", "ci_integration"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ec2b7632c5af7653ce0f946a2994e920a78fc038"}
{"id": "train_48750", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "tooling", "tests_are_truth", "auditability"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "022ab8a5f14dccbeb2685bd6cbb03e6503b33566"}
{"id": "train_48751", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "auditability", "tests_are_truth", "documentation"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b5ddaa6c5fc78a228cc03a412c43a66b4ec943e3"}
{"id": "train_48752", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "documentation", "governance", "auditability"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a49bfe1993089d1230848916435d25382622c9df"}
{"id": "train_48753", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "governance", "evaluation_metrics", "documentation"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aab1785c13f555db4314f6718fa59f40a9a691a4"}
{"id": "train_48754", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "reproducibility", "documentation", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5014eb56809eb114706ce5443909f6101d7ad443"}
{"id": "train_48755", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "tooling", "reproducibility", "security_gates"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6195cc41ddcb4355b97d5e50b4ba3c350799f4ae"}
{"id": "train_48756", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "tests_are_truth", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "659960163d75b9ebae40a89ad97129053d13323b"}
{"id": "train_48757", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "evaluation_metrics", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dc69c5fd5cfe84a83c134ee18fea0975846f0b07"}
{"id": "train_48758", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "auditability", "documentation", "governance"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1d71a047391c5fa4212185260641d03e284c1239"}
{"id": "train_48759", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "repo_scale_reasoning", "tooling", "auditability"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "28a37e8fefef59dcd395d0fcaff768e7e1b0767b"}
{"id": "train_48760", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "tooling", "cost_latency_tradeoffs", "governance"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e2c3827b087129ce7bd06834ccd89f9561852b82"}
{"id": "train_48761", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "reproducibility", "documentation", "tooling"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "275986e0fe7de35b4e0853dcecb374eac0f5d153"}
{"id": "train_48762", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "governance", "ci_integration"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "244a3e9c8b7125d64500abc1aaeb7a77f73687de"}
{"id": "train_48763", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "repo_scale_reasoning", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3b81f5f390785d82e92eed404cacda1ef7c808bd"}
{"id": "train_48764", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "security_gates", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "111baa041f307ad92fe00a2bd2c8ef58cc074b76"}
{"id": "train_48765", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "tests_are_truth", "ci_integration", "governance"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "61d2305d2e5b9b6cafc7d04d11395e7e30607bbf"}
{"id": "train_48766", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "governance", "reproducibility", "ci_integration"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8c4aa5be41d96c8593f8955eac3093402def6a8b"}
{"id": "train_48767", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "governance", "tests_are_truth", "documentation"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cd2ca39e588e9a7aa27fc2268acf6692963de60c"}
{"id": "train_48768", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "governance", "auditability", "tests_are_truth"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "001c35cf9757b9f6025b0d40d28ddeed1f04335e"}
{"id": "train_48769", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "documentation", "ci_integration", "security_gates"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b816f638eb13275815128b5016dd595be1ffd932"}
{"id": "train_48770", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["governance", "cost_latency_tradeoffs", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eebeb89bfaa6bfb01f6956937e2d8bb167bcd63a"}
{"id": "train_48771", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "tests_are_truth", "security_gates", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "05af232cfa91edfb290e8ea80df7f61079625e61"}
{"id": "train_48772", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "security_gates", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8244be2c74c47c9481618f17a570e73c31ba2892"}
{"id": "train_48773", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["repo_scale_reasoning", "documentation", "governance", "tooling"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6c071d8963d9e69daa97917228594b3d8fca1dc4"}
{"id": "train_48774", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "documentation", "tooling", "tests_are_truth"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9e8340b7c056f9fb55a075b9ed42859c5e966591"}
{"id": "train_48775", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "reproducibility", "evaluation_metrics", "documentation"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c80667e079bb4c4dfcebbca7249640a30eee3fbb"}
{"id": "train_48776", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "governance", "tests_are_truth", "auditability"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c6cbcb1a60806d0d3d4ca7a9829e38ad04b3c41e"}
{"id": "train_48777", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "tooling", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "700fd55535058ced437597a7d12c17ac86f31c6e"}
{"id": "train_48778", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "repo_scale_reasoning", "tests_are_truth", "governance"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dd7ec16e575433047ebc4c5212f4f7590967c779"}
{"id": "train_48779", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "documentation", "cost_latency_tradeoffs", "governance"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "57423224c023b38d94062e7dd0c5e0d09e396327"}
{"id": "train_48780", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c98fe176a7ddb044fbc613675cee5064e2095746"}
{"id": "train_48781", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "tests_are_truth", "tooling", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9cdf04e3b1430901483cc2a17a69912395f5571a"}
{"id": "train_48782", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "tests_are_truth", "security_gates", "documentation"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9c123e9d9c88cfd08cb42906887b701430990594"}
{"id": "train_48783", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "governance", "repo_scale_reasoning", "security_gates"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a9567731fbe9c6ba02d8ee3ae091b08fcb8220cb"}
{"id": "train_48784", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["documentation", "security_gates", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f91e3472ea5f7dc6f00cef0f239d8f08b04db2d3"}
{"id": "train_48785", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "reproducibility", "security_gates", "tooling"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d03d62cc1bf48ffbe99d0d0e3e5027575d3f6e45"}
{"id": "train_48786", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "governance", "reproducibility", "documentation"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d2eecb9547f2d10e9c30fbd8ae1c56265bffaf26"}
{"id": "train_48787", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "repo_scale_reasoning", "security_gates", "tooling"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6e21dc9bee992a3d23a46aca0127d9d9a2255842"}
{"id": "train_48788", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "ci_integration", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3e3b1562d2603259ba035767c5e0695a515f4298"}
{"id": "train_48789", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "auditability", "ci_integration", "tooling"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b3fe9149b4d03fd11c928d32578059dc18cfab90"}
{"id": "train_48790", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "governance", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4dd0264d94027a235ed4c65d222af0f299f99571"}
{"id": "train_48791", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "governance", "repo_scale_reasoning", "documentation"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "817cf340ff825502c122a2be45913aed7aca1cc5"}
{"id": "train_48792", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "reproducibility", "ci_integration", "documentation"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7581fb642a66aea84b3e47ff12f065277a1fbf29"}
{"id": "train_48793", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "documentation", "repo_scale_reasoning", "security_gates"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "230a0a9c311863f25edc1d841cad71e51eeffc74"}
{"id": "train_48794", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["governance", "evaluation_metrics", "security_gates", "documentation"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3afafc4c1286e601b3e34833a072d94c285f645a"}
{"id": "train_48795", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "security_gates", "auditability"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e7d99fc793a1c140a82bf6bfdd55631b7da4fbff"}
{"id": "train_48796", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "governance", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "edfaedf4c8126807bce1c896fc7bca42be5ad4b7"}
{"id": "train_48797", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "documentation", "reproducibility", "governance"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5756608c4df62181cb8c4c5454b5c2d02c98e2dd"}
{"id": "train_48798", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "security_gates", "tooling", "governance"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "78a4aa99c55faf3898bdcdcd8d467e6ed7aa0fd8"}
{"id": "train_48799", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "tooling", "governance", "auditability"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "500053cea39fa7057c14f56c10cfbcdee84e6d98"}
{"id": "train_48800", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["governance", "tooling", "reproducibility", "auditability"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cbc09dc7674ab8737b66279b112ae1388c718493"}
{"id": "train_48801", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "documentation", "governance", "evaluation_metrics"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e6bcf062f84e3c9890f27821c55a1e808a33690d"}
{"id": "train_48802", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "governance", "tooling", "tests_are_truth"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "504d7a943f184a5635cfcc78f1e2babc010b6541"}
{"id": "train_48803", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "tooling", "documentation", "auditability"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "48e14be1c2135d0145eee3ebacf2b713a337e4e9"}
{"id": "train_48804", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "security_gates", "ci_integration", "governance"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "247edf034544159c33ea8cc271bc92ead15754f6"}
{"id": "train_48805", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "auditability", "reproducibility", "ci_integration"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "af7e551ad262c329ad55293807f6e51fd9314a8f"}
{"id": "train_48806", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "repo_scale_reasoning", "reproducibility", "tests_are_truth"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2fbc4765fa3c2153d74d53cdd5f4565410809356"}
{"id": "train_48807", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "tooling", "auditability"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "601a8db1bf5f29e7951c8efed104870bcfdc0286"}
{"id": "train_48808", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "tests_are_truth", "ci_integration", "governance"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8c70897e144508fb77980fec13297ee795f6ae1d"}
{"id": "train_48809", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "governance", "reproducibility", "tooling"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d8f02c1f4d932a347301ba3bef48ff87e71da238"}
{"id": "train_48810", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "auditability", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a1ff4289366ab49b380e5f80de71ef692911cb15"}
{"id": "train_48811", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "tests_are_truth", "ci_integration", "governance"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2abaff69fc3f38cbb02470fb20a44710a748b2cf"}
{"id": "train_48812", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "governance", "tests_are_truth", "tooling"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4d963016ee80bab3c7cd82eb1b5cab04c06b380d"}
{"id": "train_48813", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "evaluation_metrics", "auditability", "tooling"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1c97ed724e93ef6115d5b1d88c26cab73e5b1c09"}
{"id": "train_48814", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["cost_latency_tradeoffs", "tooling", "ci_integration", "reproducibility"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b85e533e1c5326612e7b9127621a4956977c4592"}
{"id": "train_48815", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "tests_are_truth", "auditability", "governance"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "102c9f6e44d378faf9e5e9f515813f73defefd45"}
{"id": "train_48816", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "repo_scale_reasoning", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f4fe0986c87372f51cea0ad78d2014ac6ff5aff6"}
{"id": "train_48817", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "governance", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6b8f85dae81fa224cd9e36fb26c9dbc3125c3040"}
{"id": "train_48818", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["documentation", "tests_are_truth", "ci_integration", "reproducibility"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "088ae1f2a8e7c90dfe19b8a54f2f0c2f7da12245"}
{"id": "train_48819", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "ci_integration", "reproducibility", "documentation"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1d2978aa5e18f15e6e368caf8a6aac010f77a16c"}
{"id": "train_48820", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "cost_latency_tradeoffs", "documentation", "ci_integration"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0402ea2401676b6c4177a4c0170a096b191732de"}
{"id": "train_48821", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "auditability", "ci_integration", "tests_are_truth"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "14c0bdd10cf435e4f789d1cd96068398a9b8c856"}
{"id": "train_48822", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "auditability", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5f2d2a75c861e61af211b2f132f68cae1eb06441"}
{"id": "train_48823", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "security_gates", "ci_integration", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4d25a3c21c865391fe7f4ae1de28c11a4b8d4458"}
{"id": "train_48824", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "repo_scale_reasoning", "cost_latency_tradeoffs", "evaluation_metrics"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ea5e74a9e8dfdb14be0b309bde7b5e639711fb2b"}
{"id": "train_48825", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "tooling", "tests_are_truth", "auditability"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ffed4af49f0f3d6c21e6ae4ac0e3109853f578c6"}
{"id": "train_48826", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "governance", "ci_integration", "tests_are_truth"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f058e1badcd4bbc176b9392029c6a4104ab9cc8f"}
{"id": "train_48827", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "security_gates", "tooling"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b501f3dc2fbba7b0fae50a7f638fc63ac81c3bdf"}
{"id": "train_48828", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "auditability", "evaluation_metrics", "security_gates"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f6b65663886f7204efe3c351ac8a3635bd0a9528"}
{"id": "train_48829", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "documentation", "governance", "security_gates"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f5ca31cd3bc43de20aaff70be64426c970d7a8d1"}
{"id": "train_48830", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "repo_scale_reasoning", "documentation", "auditability"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fb4eb35266139c98bfaf952074468d395df199e8"}
{"id": "train_48831", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["evaluation_metrics", "tests_are_truth", "repo_scale_reasoning", "documentation"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0bf037373edfaadeeb9bd53cb672cb1c10945f36"}
{"id": "train_48832", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8e72d76c8188fd3135b23d89f658c80e54827520"}
{"id": "train_48833", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation and policy gates\nDifficulty: expert\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["auditability", "tooling", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "50aa309132bfbc50b31b812480a11c3a45b6e4f1"}
{"id": "train_48834", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "auditability", "security_gates"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ed25fee20c73ecd2047815d9349e8c61ef114205"}
{"id": "train_48835", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "auditability", "tests_are_truth", "governance"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "55a869b6853401241f18c95dc0fe0703306da9d5"}
{"id": "train_48836", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "security_gates", "governance"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "463d76314ed7bfd0bca06f3535ecd5dba2e2bd2b"}
{"id": "train_48837", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "governance", "auditability", "ci_integration"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e50c942bc705d01435dbe78b20fa917f8971768c"}
{"id": "train_48838", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "governance", "evaluation_metrics", "ci_integration"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b85c2acdd7f3d0130add9b64af2931695bb0d9b7"}
{"id": "train_48839", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "governance", "reproducibility", "auditability"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a06b9eea8584902a63ff9151d9a695570b60ed35"}
{"id": "train_48840", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tests_are_truth", "tooling", "ci_integration", "evaluation_metrics"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f91a09497b621d12dbf2e24ba35ba39305a5e8bc"}
{"id": "train_48841", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tests_are_truth", "security_gates", "evaluation_metrics", "governance"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8702533e65c9e003e67116834e63bbde136fe94d"}
{"id": "train_48842", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "evaluation_metrics", "documentation", "tooling"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "64e4571618e0986c4dbbbc799e1e676739bbadaa"}
{"id": "train_48843", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "governance", "tooling", "auditability"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f413e73d7821a0d3e2557a3e4d8b029e21a7224e"}
{"id": "train_48844", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "tests_are_truth", "tooling", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bd5eb9a3752e6814e823eb37c602ddfc36180bb7"}
{"id": "train_48845", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "cost_latency_tradeoffs", "auditability", "governance"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dd0a8be38449b3cd4c655c7979f9735ae0aabdeb"}
{"id": "train_48846", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: JavaScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["tooling", "reproducibility", "evaluation_metrics", "tests_are_truth"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0bb9efa042c3bb4b5af881944f4535ae340cde24"}
{"id": "train_48847", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "tests_are_truth", "governance", "reproducibility"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "33c8d3af2e98e543c3a8bc600705963730549e40"}
{"id": "train_48848", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["auditability", "repo_scale_reasoning", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "724939db5db965610acc40faa44ca68a894eadb2"}
{"id": "train_48849", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "governance", "tests_are_truth", "reproducibility"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3e1fa0a19adead1adfe9198f6743240fa9f74614"}
{"id": "train_48850", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "cost_latency_tradeoffs", "governance", "tooling"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6945b99adaeefaff7a0ee96224c7f38853bae2ba"}
{"id": "train_48851", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "tests_are_truth", "repo_scale_reasoning", "evaluation_metrics"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fadc3664022d793f1eadeeb576780fdc1c420948"}
{"id": "train_48852", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["governance", "documentation", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8a442d2000f833f6e415b225ef5c2a5ec8af7cf9"}
{"id": "train_48853", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "failure_analysis", "difficulty": "intermediate", "instruction": "Task: failure_analysis\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "tests_are_truth", "documentation", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f5a8c55e868e6387f8468a56ddb60172fc97d8bd"}
{"id": "train_48854", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "tooling", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "682cbb0da9499292695dac98e68c836808a3bc41"}
{"id": "train_48855", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "tests_are_truth", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8b56f0c4f71c4c86a35a408e3080161e22dda34c"}
{"id": "train_48856", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "evaluation_metrics", "security_gates", "auditability"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7632bca8831d1b73b293b27090ab30705ff75683"}
{"id": "train_48857", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["evaluation_metrics", "auditability", "reproducibility", "documentation"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2910c0cd8e9ce0a987b7426743a9bee0048e5308"}
{"id": "train_48858", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "ci_integration", "tooling"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cf580dd9d2084bd2be9c0b5d413267e3b954a097"}
{"id": "train_48859", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "evaluation_metrics", "documentation", "tooling"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7e0f35071e988f91cfb5201d9c15e87ae4832c25"}
{"id": "train_48860", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: intermediate\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["security_gates", "tests_are_truth", "documentation", "tooling"], "moe_experts": ["evaluation_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c5a7085bbd8e0063738302792f2b0d6b83e9a50c"}
{"id": "train_48861", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "ci_integration", "reproducibility", "governance"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "dc118fcf2bc4ce0522c174b4aa0060e9ff95331d"}
{"id": "train_48862", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "auditability", "reproducibility", "security_gates"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d24ddbd475cb7bf8b3e680269dc96028463de666"}
{"id": "train_48863", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["security_gates", "repo_scale_reasoning", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3f1950202571672404226cb4b67328328074e271"}
{"id": "train_48864", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["tooling", "ci_integration", "reproducibility", "governance"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c622bc869fdeb177437f2491353e633cf9343858"}
{"id": "train_48865", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["security_gates", "governance", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "df2b3f46616fa0c47cd46cc1109a4eab1d07e264"}
{"id": "train_48866", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "tooling", "documentation"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4b3b39d29d50323687f31cdd02ba47dc558580b9"}
{"id": "train_48867", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "tests_are_truth", "security_gates"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b2740e998ae6d8483d6deb036de9a7be35df20d0"}
{"id": "train_48868", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "governance", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "449e8aa7daffdcd92c4a729c9dc89643eab1cd0c"}
{"id": "train_48869", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["documentation", "reproducibility", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9880a9941d0661b112f7ae785d9c5d545eac6b7d"}
{"id": "train_48870", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "governance", "reproducibility", "tests_are_truth"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0101aea55700892f232170a96b53e48868b1741c"}
{"id": "train_48871", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "evaluation_metrics", "reproducibility", "governance"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c08487bac94438075fc5cb279e718cdfc84449f5"}
{"id": "train_48872", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["governance", "evaluation_metrics", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "23847aec4837cdae28d64bab74fe4ae4ec526569"}
{"id": "train_48873", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["security_gates", "tooling", "ci_integration", "governance"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "77ed8ff11481e12ae6128fce11433eb867917a64"}
{"id": "train_48874", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "repo_scale_reasoning", "documentation", "auditability"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "eb54ef4a3198491078d12b257eadae67f68ae3a6"}
{"id": "train_48875", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "cost_latency_tradeoffs", "auditability", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a1ad92aae4ef4d7585829795917a4f2ad9b5be9c"}
{"id": "train_48876", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "governance", "repo_scale_reasoning", "tooling"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1fbbedef2e55f3fa5b8d85b7b8eb97af0289c53f"}
{"id": "train_48877", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "tests_are_truth", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ee9c2efa0cad28cc1c8d8036bcfc956254ce75ce"}
{"id": "train_48878", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "auditability", "cost_latency_tradeoffs", "reproducibility"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5979ef1aa601412977b61c3eb94b8f0cefadb858"}
{"id": "train_48879", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "ci_integration", "evaluation_metrics", "auditability"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a2add859e2ee21c2566fe10559defe678f40c8f0"}
{"id": "train_48880", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "repo_scale_reasoning", "auditability", "tests_are_truth"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "cd04c65643dbb0b8ccdbc0200279f05d6e77943d"}
{"id": "train_48881", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "evaluation_metrics", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3a6fc290633ee6e79c00f06e43776c2c4ad5b936"}
{"id": "train_48882", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "reproducibility", "repo_scale_reasoning", "governance"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "948daeb0f7d9474a690854bb0224695261033fd6"}
{"id": "train_48883", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["auditability", "reproducibility", "repo_scale_reasoning", "ci_integration"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "09d2d0194f6c36d1b5604d4a35ff45f00cf19228"}
{"id": "train_48884", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Rust\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["documentation", "evaluation_metrics", "auditability", "security_gates"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4f9dc402862a959424abd596ffb1178da44c38fc"}
{"id": "train_48885", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "auditability", "security_gates", "ci_integration"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0ad3c82b61f50932ddc34e36f96a055c053ad734"}
{"id": "train_48886", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["reproducibility", "documentation", "governance", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4312f8c9026018b9aca6e3f16a5dbdb5b5d9f351"}
{"id": "train_48887", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "tooling", "governance", "auditability"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "052ef3ec5936d5b5a2e97cd43d7c2ae4cc363905"}
{"id": "train_48888", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["auditability", "cost_latency_tradeoffs", "tooling", "evaluation_metrics"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5367ad6b3eb2718bc5e4c293238f97e00c4f79bc"}
{"id": "train_48889", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["repo_scale_reasoning", "tests_are_truth", "documentation", "tooling"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c11127ea99ba814de6cc48d62197c0eb15e06e94"}
{"id": "train_48890", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["ci_integration", "auditability", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a4dd36662148c54c4e3d0f4b8ee3789e60187b72"}
{"id": "train_48891", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "tooling", "evaluation_metrics", "governance"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fcc623e3139458c127b99c2dd4158252e6585d6b"}
{"id": "train_48892", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "evaluation_metrics", "repo_scale_reasoning", "tests_are_truth"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "258bce57b732aeb3ca53ebee74c8168a8cdf699d"}
{"id": "train_48893", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "repo_scale_reasoning", "tests_are_truth", "governance"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "42cd0531e9e8eff3e3d6210d1c7b1d1e6c50c304"}
{"id": "train_48894", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "documentation", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "53c78dd427b90de500938e9d5d064889ccbfafce"}
{"id": "train_48895", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "repo_scale_reasoning", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9357ab97ec96cee1e0801bf50cf85d26d2f9a210"}
{"id": "train_48896", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["reproducibility", "cost_latency_tradeoffs", "documentation", "security_gates"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "609fb123da95bb02881c87d4f3fa32fc7373a3db"}
{"id": "train_48897", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "evaluation_metrics", "cost_latency_tradeoffs", "governance"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7b23f6a8665c5486f9e72053065d8b82af75894e"}
{"id": "train_48898", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "auditability", "governance", "security_gates"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a4e6cd030c360ecf99d9da01cbd11033e5c4c3ec"}
{"id": "train_48899", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["tests_are_truth", "evaluation_metrics", "tooling", "repo_scale_reasoning"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2a0ddea746733ad0a9cdd371821f257b19c813b8"}
{"id": "train_48900", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["repo_scale_reasoning", "tooling", "tests_are_truth", "security_gates"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4e5f0dd2d556b6f6eed0b06d5ef577e48a30d306"}
{"id": "train_48901", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["tests_are_truth", "documentation", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f5bb48c67a8b6b578c1961d6dc4e7581a0f51666"}
{"id": "train_48902", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "reproducibility", "auditability"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "343394c7533375e2a57a6a4c8c1df5be42a675a7"}
{"id": "train_48903", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "governance", "repo_scale_reasoning", "documentation"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b013c9b886c5fd90f4a9efd1b0f7ad88fcc51de3"}
{"id": "train_48904", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "auditability", "ci_integration", "governance"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3789740d6ad21228e75e7ba29797474437549b7c"}
{"id": "train_48905", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "governance", "ci_integration", "reproducibility"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc9d581b8f4fd479a93f412f580aa10ce945cdb3"}
{"id": "train_48906", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "auditability", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d4ae3317f4b12c9082921bdd9e73b639435f84d3"}
{"id": "train_48907", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["governance", "security_gates", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b782516573836832a46dc84d630b4772bcba9b2b"}
{"id": "train_48908", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["tooling", "ci_integration", "cost_latency_tradeoffs", "tests_are_truth"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b9f11264d9a489871d909b70182a044d70f9de00"}
{"id": "train_48909", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "tooling", "reproducibility", "governance"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "84d899f048aeab510664989f128bb871802ce1ae"}
{"id": "train_48910", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "security_gates", "governance", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "4331de68332cecd2ea0f0cc2d0f5b7ded1b17841"}
{"id": "train_48911", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "reproducibility", "governance", "evaluation_metrics"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ec698cc5d6afe1289af0473c52971ad2ddd96899"}
{"id": "train_48912", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["ci_integration", "auditability", "reproducibility", "repo_scale_reasoning"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "21481491ec186e00c4335b94598f630afa8eb992"}
{"id": "train_48913", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "repo_scale_reasoning", "security_gates"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1813944b51592df998a13a5f4061d380821aa378"}
{"id": "train_48914", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["tests_are_truth", "documentation", "ci_integration", "governance"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7d5a6809087d9e93c337985babdce4e714e58c8f"}
{"id": "train_48915", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "ci_integration", "reproducibility", "documentation"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d6592e54f54b2f43c740e4608ecadf754ccfa178"}
{"id": "train_48916", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "tooling", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "faa0816d9ad9b03d929c4a7fd35e75555652e831"}
{"id": "train_48917", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "security_gates", "ci_integration"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "86cad0b0cfb8f71735785306ceaf48394e3e2ed5"}
{"id": "train_48918", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "tooling", "ci_integration", "governance"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c7ed5cf9c0c084f399d7eafc9749d5de38ce321c"}
{"id": "train_48919", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["repo_scale_reasoning", "documentation", "tooling", "tests_are_truth"], "moe_experts": ["coding_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6bcde5cbc8021aa664b9e71b5bec113bd611f837"}
{"id": "train_48920", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["governance", "cost_latency_tradeoffs", "tooling", "tests_are_truth"], "moe_experts": ["governance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "fe3e926ba5abaef629d0bd825aad9d08dd81470c"}
{"id": "train_48921", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "security_gates", "reproducibility", "tooling"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2daddfb16c571bb8505e1e554629bbdcc8fa7e6c"}
{"id": "train_48922", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "cost_latency_tradeoffs", "tooling", "evaluation_metrics"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ec023554f72bc7a11533baaec74f01f76c721be7"}
{"id": "train_48923", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "reproducibility", "auditability"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7a9b2f912f13f560ca14c9fb9778c23efc73e4c3"}
{"id": "train_48924", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "reproducibility", "ci_integration", "tooling"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b5dc8a8b7f1707c86c4646b39f8a306ca2016577"}
{"id": "train_48925", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["documentation", "governance", "auditability", "evaluation_metrics"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d7e79ba9d9c225f76b8cb6b6a2fa0cd7cb748f0f"}
{"id": "train_48926", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tooling", "auditability", "evaluation_metrics", "reproducibility"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d772a6290b841a6b4b4882b6058b5c5fb048a425"}
{"id": "train_48927", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "evaluation_metrics", "ci_integration", "reproducibility"], "moe_experts": ["agentic_systems_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "77fc10250ec5a135cef29768a6fbad4658e031f4"}
{"id": "train_48928", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: SQL\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "cost_latency_tradeoffs", "tooling", "security_gates"], "moe_experts": ["security_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bea23bd133244b16b8899bb89e01b7b02cc84366"}
{"id": "train_48929", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: C#\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "documentation", "evaluation_metrics", "cost_latency_tradeoffs"], "moe_experts": ["coding_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "972e45e41c3a1b3886f608b967696026a26c76f8"}
{"id": "train_48930", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "auditability", "tooling", "ci_integration"], "moe_experts": ["coding_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f171040221edc5cc584ef9a066c3ccb4cb5395da"}
{"id": "train_48931", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["evaluation_metrics", "documentation", "auditability", "tooling"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c1be034040e42415513c937a87f87309ca772139"}
{"id": "train_48932", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["auditability", "evaluation_metrics", "tests_are_truth", "ci_integration"], "moe_experts": ["data_curation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "5ac2fdf94c7ad304df174bf565f9cc9dafad4fa7"}
{"id": "train_48933", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "auditability", "reproducibility", "security_gates"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f369e89481502db537bd26cd19312952bcb4bb65"}
{"id": "train_48934", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["ci_integration", "auditability", "governance", "security_gates"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "be745b92c6e8a20883b195c926ffa86060f0fc91"}
{"id": "train_48935", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "ci_integration", "governance", "tooling"], "moe_experts": ["agentic_systems_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7a9aa53c7fc57fcaae65bbb8a5fabeaa2c4fb331"}
{"id": "train_48936", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Rust\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["security_gates", "repo_scale_reasoning", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b131d451e043fa94e9c9cc67bd08e85d43e7ed0a"}
{"id": "train_48937", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["documentation", "tooling", "ci_integration", "security_gates"], "moe_experts": ["governance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f1d113c30976ec3a9d03fa6690ae6c55bc77315e"}
{"id": "train_48938", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Java\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["tests_are_truth", "security_gates", "ci_integration", "governance"], "moe_experts": ["performance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c5661dbc5a47e63f243ec8e9f4476925867f657d"}
{"id": "train_48939", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["reproducibility", "security_gates", "tests_are_truth", "evaluation_metrics"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9b8727937fc2f7abb8533d971ba41ae703dd20f8"}
{"id": "train_48940", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "agent_loop", "difficulty": "intermediate", "instruction": "Task: agent_loop\nTopic: Self-improving agents and feedback loops\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "tests_are_truth", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "80c077d6df94f9f66a2bee02f8b4e04b2973a480"}
{"id": "train_48941", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["evaluation_metrics", "documentation", "security_gates", "reproducibility"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ea4e116f17623d329d2fd6b42da6a746c6123006"}
{"id": "train_48942", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "repo_scale_reasoning", "reproducibility", "evaluation_metrics"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e4eab50b57e041b57cd2841d61482be6508e669f"}
{"id": "train_48943", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "documentation", "reproducibility", "security_gates"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "75065abdef662044082065e36be9e1b3924cfe5b"}
{"id": "train_48944", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: JavaScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["ci_integration", "reproducibility", "security_gates", "cost_latency_tradeoffs"], "moe_experts": ["data_curation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f28d9679ccdcc9902ee4de64760c0a792dc1a02f"}
{"id": "train_48945", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["governance", "ci_integration", "repo_scale_reasoning", "reproducibility"], "moe_experts": ["security_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "653a1ee7521fc2f6c8bcd3764ec9a6c5c48d1849"}
{"id": "train_48946", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: SQL\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["documentation", "repo_scale_reasoning", "governance", "auditability"], "moe_experts": ["performance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "73acb5d6ea44d01e671482abb996240ab37b2fdf"}
{"id": "train_48947", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["cost_latency_tradeoffs", "auditability", "repo_scale_reasoning", "documentation"], "moe_experts": ["security_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "358e5a8f49e9c52a1eec7d00f91835f4400b7dc8"}
{"id": "train_48948", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["documentation", "reproducibility", "security_gates", "ci_integration"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bb1a98b7ca24f295fed3236dadf4561b297b60e0"}
{"id": "train_48949", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and licensing for code data\nDifficulty: expert\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "evaluation_metrics", "reproducibility", "governance"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2c56237399e36a54cab96f63d4249e0d24414929"}
{"id": "train_48950", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "governance", "auditability", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "2bf52e9f8d52ed2b8f6a67f57d44676ae76b0d2d"}
{"id": "train_48951", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["security_gates", "ci_integration", "reproducibility", "cost_latency_tradeoffs"], "moe_experts": ["performance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9149e7a0502d8a0626e7461feb1d5934f24615bc"}
{"id": "train_48952", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "tests_are_truth", "repo_scale_reasoning", "tooling"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7e0f35071e988f91cfb5201d9c15e87ae4832c25"}
{"id": "train_48953", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: expert\nTarget language: Java\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "ci_integration", "tests_are_truth", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "20c5c0c501d9723841dde21892ebc1717d11a5e9"}
{"id": "train_48954", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["governance", "cost_latency_tradeoffs", "security_gates", "documentation"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "695cc85b8c2fb29fcbc3f01e76fb96c516050ace"}
{"id": "train_48955", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: advanced\nTarget language: JavaScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["cost_latency_tradeoffs", "tests_are_truth", "documentation", "governance"], "moe_experts": ["evaluation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a4769f12fe446f9619c88852a483fa91f774c620"}
{"id": "train_48956", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: intermediate\nTarget language: C#\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["evaluation_metrics", "documentation", "tests_are_truth", "cost_latency_tradeoffs"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "adf25af28747d5e524d119ea3876201befa1ec8d"}
{"id": "train_48957", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: advanced\nTarget language: TypeScript\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "documentation", "ci_integration", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9648fe749ce74968945762ac7559a00b68e1bf73"}
{"id": "train_48958", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["governance", "tooling", "evaluation_metrics", "auditability"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1e64cec33eb13e61bce3e09c0b3bee1cf72fe559"}
{"id": "train_48959", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["reproducibility", "evaluation_metrics", "repo_scale_reasoning", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c0b73e1c4722443f07da0f5ecbd5d1dd9371ca7e"}
{"id": "train_48960", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Bash\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["cost_latency_tradeoffs", "auditability", "documentation", "security_gates"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "23424c81e0d8c4da2c5d6f0c3c0be2f9fc39590e"}
{"id": "train_48961", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context and repo-scale understanding\nDifficulty: advanced\nTarget language: Go\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "auditability", "reproducibility", "documentation"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6ec572c67606f4eb5a76bc2386f2516ef5d67f67"}
{"id": "train_48962", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["governance", "documentation", "cost_latency_tradeoffs", "security_gates"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a1904b3c6f15ef3e330f9ecf27ec51f41a605cf8"}
{"id": "train_48963", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "cost_latency_tradeoffs", "security_gates", "tests_are_truth"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9a54df753b4e97085ac69cea6245678164f02132"}
{"id": "train_48964", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["tooling", "evaluation_metrics", "ci_integration", "repo_scale_reasoning"], "moe_experts": ["evaluation_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c1a57b97529ff0aa96e152049b448d02198cdba6"}
{"id": "train_48965", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["tooling", "cost_latency_tradeoffs", "reproducibility", "tests_are_truth"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7098555f96702cb4007c3275d102183f456cc94f"}
{"id": "train_48966", "created": "2026-01-01", "topic": "Self-improving agents and feedback loops", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Self-improving agents and feedback loops\nDifficulty: expert\nTarget language: Bash\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "ci_integration", "governance", "evaluation_metrics"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c4d5c38e7f7ecb7b9131416143da8bb3911a3eb7"}
{"id": "train_48967", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["ci_integration", "documentation", "evaluation_metrics", "security_gates"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7c01fa930285603dd534b0b5eb8187b3a646dea0"}
{"id": "train_48968", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nTarget language: C#\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["security_gates", "reproducibility", "documentation", "ci_integration"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "aaf1293581a52633b7c2a147a759b37058848abf"}
{"id": "train_48969", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: expert\nTarget language: Java\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["cost_latency_tradeoffs", "evaluation_metrics", "auditability", "ci_integration"], "moe_experts": ["evaluation_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "73d06da98b55a7502cfa0098e28e446de25c7e18"}
{"id": "train_48970", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "tooling", "documentation", "repo_scale_reasoning"], "moe_experts": ["governance_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bd2656a0e6bc0a5153184824fe4af848774be1cd"}
{"id": "train_48971", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: intermediate\nTarget language: Python\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["governance", "evaluation_metrics", "security_gates", "tests_are_truth"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "768ae4c0b367829d5b635d28fd5a56a1721f678c"}
{"id": "train_48972", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: C#\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "C#", "developer_needs": ["reproducibility", "evaluation_metrics", "tests_are_truth", "documentation"], "moe_experts": ["performance_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f9e6dd372d6f303439310902eb2a9855a31c3644"}
{"id": "train_48973", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: expert\nTarget language: Python\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nScaffold:\n```python\ndef agent_loop(plan, edit, test, issue, max_iters=4):\n    history = []\n    p = plan(issue)\n    for _ in range(max_iters):\n        patch = edit(issue, p)\n        ok, report = test(patch)\n        history.append({\"plan\": p, \"ok\": ok})\n        if ok:\n            return patch, history\n        p = p + \" | refine\"\n    return patch, history\n\n```\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["governance", "cost_latency_tradeoffs", "reproducibility", "tests_are_truth"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "3c19a33369b9146cb38c50bf05780518d79aa2e4"}
{"id": "train_48974", "created": "2026-01-01", "topic": "Dataset curation pipelines (filter, dedupe, quality)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines (filter, dedupe, quality)\nDifficulty: advanced\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "documentation", "tooling", "ci_integration"], "moe_experts": ["evaluation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e729f89816dce29fff86a4cc8980dd7438e072a8"}
{"id": "train_48975", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "code", "difficulty": "intermediate", "instruction": "Task: code\nTopic: Extended context and repo-scale understanding\nDifficulty: intermediate\nTarget language: C#\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "C#", "developer_needs": ["documentation", "auditability", "cost_latency_tradeoffs", "ci_integration"], "moe_experts": ["data_curation_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d4fb713b6b246d7e7a8fe267a4924cb093acc86e"}
{"id": "train_48976", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "compare", "difficulty": "intermediate", "instruction": "Task: compare\nTopic: SWE-bench style real-repo evaluation\nDifficulty: intermediate\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nCompare: capability, cost, latency, reliability\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["auditability", "repo_scale_reasoning", "cost_latency_tradeoffs", "tooling"], "moe_experts": ["performance_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d5e2aaff5d2f68b652695d95269fe281bc286d0e"}
{"id": "train_48977", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench style real-repo evaluation\nDifficulty: advanced\nTarget language: TypeScript\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "TypeScript", "developer_needs": ["ci_integration", "evaluation_metrics", "security_gates", "governance"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8a87daa32bd1f37fad71cf767e4a5e9d2c079400"}
{"id": "train_48978", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "intermediate", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: intermediate\nTarget language: Rust\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["evaluation_metrics", "reproducibility", "tooling", "repo_scale_reasoning"], "moe_experts": ["agentic_systems_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "ac73df952c088e4ee7b542f6b70bc619dcca3076"}
{"id": "train_48979", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Java\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": [], "metadata": {"target_language": "Java", "developer_needs": ["repo_scale_reasoning", "governance", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8de0e11e91e4483ab3b1e2c5c92c30eb8d63e07b"}
{"id": "train_48980", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, and reliability optimization\nDifficulty: advanced\nTarget language: Python\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["repo_scale_reasoning", "tooling", "cost_latency_tradeoffs", "documentation"], "moe_experts": ["performance_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b3431b57e8695453fcd0ffea7bf6463c948ef17c"}
{"id": "train_48981", "created": "2026-01-01", "topic": "Multimodal dev workflows (docs, diagrams, traces)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Multimodal dev workflows (docs, diagrams, traces)\nDifficulty: advanced\nTarget language: SQL\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["governance", "repo_scale_reasoning", "documentation", "auditability"], "moe_experts": ["governance_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e3973f4af910fafa9528548e140dc2e6a3d24915"}
{"id": "train_48982", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "eval", "difficulty": "intermediate", "instruction": "Task: eval\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: intermediate\nTarget language: Bash\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nEval:\npass@k, time-to-green, regressions, diff size\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Bash", "developer_needs": ["tests_are_truth", "documentation", "auditability", "tooling"], "moe_experts": ["evaluation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "a7571c5f2978e2ddb81cc4e135f08942cca07cb6"}
{"id": "train_48983", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "patch_diff", "difficulty": "intermediate", "instruction": "Task: patch_diff\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: JavaScript\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "JavaScript", "developer_needs": ["repo_scale_reasoning", "auditability", "documentation", "tests_are_truth"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "c3bd8e610cc98a9d7b85a4d01346af84afa3c483"}
{"id": "train_48984", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, and reliability optimization\nDifficulty: expert\nTarget language: Go\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "evaluation_metrics", "security_gates", "tooling"], "moe_experts": ["agentic_systems_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "72a8777832cb7138f3070141518ca68be7714f17"}
{"id": "train_48985", "created": "2026-01-01", "topic": "Latency, cost, and reliability optimization", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Latency, cost, and reliability optimization\nDifficulty: intermediate\nTarget language: Go\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Go", "developer_needs": ["cost_latency_tradeoffs", "auditability", "evaluation_metrics", "repo_scale_reasoning"], "moe_experts": ["coding_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "6f6359dfd71813231a1369c61f0af31b2b9aa991"}
{"id": "train_48986", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: SQL\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "SQL", "developer_needs": ["tests_are_truth", "governance", "evaluation_metrics", "documentation"], "moe_experts": ["security_expert", "data_curation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "8bdd6a277ce236a8659e585f0364815b194e4d71"}
{"id": "train_48987", "created": "2026-01-01", "topic": "Secure code generation and policy gates", "task_type": "data_pipeline", "difficulty": "intermediate", "instruction": "Task: data_pipeline\nTopic: Secure code generation and policy gates\nDifficulty: intermediate\nTarget language: TypeScript\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "TypeScript", "developer_needs": ["reproducibility", "documentation", "ci_integration", "security_gates"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e257b5d4a83d9a61a30a5b00042ab3ea79cd0c3e"}
{"id": "train_48988", "created": "2026-01-01", "topic": "Tool calling, sandboxes, and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling, sandboxes, and CI integration\nDifficulty: expert\nTarget language: Rust\nContext: Research team validating claims against real repos.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPatch (diff-style):\n```diff\n- if x == 0:\n-     return 1/x\n+ if x == 0:\n+     raise ValueError('division by zero')\n```\nAcceptance:\n- Tests pass\n- No new regressions\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["cost_latency_tradeoffs", "tooling", "security_gates", "evaluation_metrics"], "moe_experts": ["coding_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "1155cf217b74f8a0a64b9561d763879209265c2f"}
{"id": "train_48989", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Java\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Java", "developer_needs": ["auditability", "ci_integration", "evaluation_metrics", "reproducibility"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "d42831b6e9bfd09597a62b1cbca030bd6cdb2614"}
{"id": "train_48990", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: advanced\nTarget language: Bash\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["evaluation_metrics", "ci_integration", "tooling", "cost_latency_tradeoffs"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "b7b74c74bf938238f5847088f13ce087af5f2e8b"}
{"id": "train_48991", "created": "2026-01-01", "topic": "Code-specialized model families and sizing tradeoffs", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and sizing tradeoffs\nDifficulty: expert\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["evaluation_metrics", "repo_scale_reasoning", "documentation", "tests_are_truth"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "649b2f8c09c54d76b4c310b1ff90ed0feda24e26"}
{"id": "train_48992", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "review", "difficulty": "intermediate", "instruction": "Task: review\nTopic: Governance, provenance, and licensing for code data\nDifficulty: intermediate\nTarget language: Go\nContext: Large monorepo with flaky tests and strict CI.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nReview: correctness, security, performance, governance\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["ci_integration", "governance", "auditability", "tooling"], "moe_experts": ["security_expert", "evaluation_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "0fc1704a31099547f974d9daabbd580673064fde"}
{"id": "train_48993", "created": "2026-01-01", "topic": "Governance, provenance, and licensing for code data", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and licensing for code data\nDifficulty: advanced\nTarget language: Go\nContext: Regulated environment requiring audit trails.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Go", "developer_needs": ["repo_scale_reasoning", "security_gates", "tooling", "tests_are_truth"], "moe_experts": ["data_curation_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "f84af4a6996eeb4976689b9186fc2407c97e196f"}
{"id": "train_48994", "created": "2026-01-01", "topic": "Model merging, distillation, and continued pretraining", "task_type": "explain", "difficulty": "intermediate", "instruction": "Task: explain\nTopic: Model merging, distillation, and continued pretraining\nDifficulty: intermediate\nTarget language: Bash\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Bash", "developer_needs": ["auditability", "repo_scale_reasoning", "evaluation_metrics", "reproducibility"], "moe_experts": ["security_expert", "agentic_systems_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "9cdf04e3b1430901483cc2a17a69912395f5571a"}
{"id": "train_48995", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: SQL\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "SQL", "developer_needs": ["security_gates", "cost_latency_tradeoffs", "governance", "evaluation_metrics"], "moe_experts": ["governance_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "e183fe22439075b02dd35e4533fd571bc6f10a73"}
{"id": "train_48996", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) for code\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nLoop: Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human gate\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tests_are_truth", "auditability", "evaluation_metrics", "tooling"], "moe_experts": ["coding_expert", "security_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "67848c1fe56f084369d97d96331063a43a347404"}
{"id": "train_48997", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nTarget language: Python\nContext: Offline/local deployment with limited compute.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nFailure:\n- Initial patch broke edge case\nReflection:\n- Missing zero-input guard\nCorrection:\n- Add explicit validation + test\n", "reasoning_steps": ["Identify failing behavior via tests", "Minimize change surface", "Apply patch", "Run targeted + full suite", "Verify no regressions"], "metadata": {"target_language": "Python", "developer_needs": ["governance", "auditability", "tests_are_truth", "tooling"], "moe_experts": ["governance_expert", "coding_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "7ed80a9a853b9346cde74d81afaa498bb55e6d96"}
{"id": "train_48998", "created": "2026-01-01", "topic": "Extended context and repo-scale understanding", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Extended context and repo-scale understanding\nDifficulty: expert\nTarget language: Rust\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nPipeline:\nIngest \u2192 Normalize \u2192 Filter \u2192 Dedupe \u2192 Quality \u2192 Mix \u2192 Audit\n", "reasoning_steps": [], "metadata": {"target_language": "Rust", "developer_needs": ["tooling", "tests_are_truth", "cost_latency_tradeoffs", "repo_scale_reasoning"], "moe_experts": ["data_curation_expert", "performance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "737d1d360e2b3419c9de6cf1cda6974b57a1e799"}
{"id": "train_48999", "created": "2026-01-01", "topic": "SWE-bench style real-repo evaluation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench style real-repo evaluation\nDifficulty: expert\nTarget language: Python\nContext: High-traffic service with latency SLOs.\nProduce expert-level, production-ready artifacts.", "input": "", "output": "Facts:\n- Modern AI coding prioritizes correctness, evaluation, and governance.\n- Agentic loops with test gates outperform single-pass generation.\n\nDesign with risks, metrics, acceptance criteria\n", "reasoning_steps": [], "metadata": {"target_language": "Python", "developer_needs": ["tooling", "tests_are_truth", "governance", "documentation"], "moe_experts": ["agentic_systems_expert", "governance_expert"], "governance": {"audit_required": true, "tests_required": true}}, "hash": "bc11e6fc8110586b0f0100862af05b6ce8971adb"}