Fine-tune → honestly evaluate on held-out data → serve in production. Fine-tuned models, synthetic datasets, and live demos by Laela Zorana.
Laela Zorana PRO
LaelaZ
·
AI & ML interests
Most ML portfolios stop at the demo. I take a model from the GPU kernel up to a real API and measure where it's wrong on data it never saw. Focus: GPU kernels (Triton/CUDA), deep learning, fine-tuning & serving, LLM eval & red-teaming, RAG & agents, MLOps, Python performance, quantization.
Recent Activity
updated a dataset about 23 hours ago
LaelaZ/vla-action-anomaly-eval published a dataset about 23 hours ago
LaelaZ/vla-action-anomaly-eval updated a Space 1 day ago
LaelaZ/embodied-efficiencyOrganizations
spaces 17
pinned
Running
embodied-efficiency
🤖
Deploy console for VLAs, compiler + safety supervisor
pinned
Sleeping
Solidity Audit AI
🛡
Rule-based Solidity static auditor + optional LLM fixes
pinned
Sleeping
InterviewCoach
🎯
AI interview prep with rubric-scored feedback
pinned
Sleeping
LLM Security Scanner
🔐
Red-team scanner for LLM apps + governance pack
pinned
Sleeping
SupportCopilot
💬
AI e-commerce support automation + ROI dashboard
pinned
Sleeping
Emotion Spectrum API
🎭
DistilBERT emotion classifier — live demo + API