Participating in the Mayo Clinic Accelerate Program. Healthcare AI startup focused on extracting clinical features from medical records.
Responsibilities:
- Deploy offline extraction LLM (open-weight, vLLM/llama.cpp)
- Train supervised audit models on clinical data
- Build evaluation harness and fairness slicing pipeline
- Prepare explainability artifacts (SHAP, model cards)
Requirements:
- 5+ years ML engineering
- Python
- Supervised tabular ML: LightGBM, XGBoost, scikit-learn
- Deployment of open-weight LLM: vLLM/llama.cpp, quantization (GGUF/AWQ/GPTQ)
- Experience in regulated/air-gapped environments
- Evaluation rigor: stratified hold-outs, bootstrap CIs, SHAP
Optional: Knowledge of clinical NLP, US ICD-10/CPT, Fairlearn/AIF360.