aydin237/medbrain-squad-scribe

MedBrainSquad — Scribe Agent Adapter

LoRA fine-tuned adapter for MedGemma 1.5 4B, trained to convert raw clinical shorthand and doctor dictation into structured SOAP JSON format.

Project

Part of the MedBrainSquad Clinical Safety Net — a three-tier adversarial multi-agent pipeline for safe, hallucination-resistant clinical documentation.

🔗 Full project: github.com/aydiny/medbrain-squad 🔗 Auditor Adapter: aydin237/medbrain-squad-auditor

Training Details

• Base model: google/medgemma-1.5-4b-it
• Method: LoRA (Rank 16, Alpha 32)
• Dataset: 500-row Teacher-Student distillation dataset (Gemini 3.0 Pro teacher → MedGemma student)
• Hardware: Dual NVIDIA T4 (Kaggle)
• Framework: HuggingFace PEFT + TRL

Intended Use

Research and development of clinical NLP pipelines. Demonstrates adversarial multi-agent safety architecture for regulated healthcare environments (NHS/EU GDPR).

⚠️ Limitations & Disclaimer

• Research prototype — not validated for clinical deployment
• Base model (MedGemma 1.5 4B) subject to Google HAID Terms
• Known hallucination risk in edge cases — mitigated by Auditor agent in the full pipeline
• Trained on synthetic data only