Description

Key responsibilities

• Define the LoRA fine-tuning strategy — target tasks, data selection, prompt/label design, and success metrics — against the frozen Foundation base model.
• Curate, clean, and structure the clinical training corpus; design train/validation/test splits and guard against leakage.
• Design and run the evaluation harness: accuracy and clinical-quality metrics, baseline-vs-adapter comparisons, error analysis, and bias/safety checks.
• Tune hyperparameters and iterate on data/approach based on evaluation results.
• Partner with ML Engineers on experiment setup and with domain/clinical SMEs to validate outputs against medical ground truth.
• Document methodology, assumptions, and model-quality findings for stakeholders.

Required skills: Python, PyTorch, Hugging Face (Transformers/PEFT), LLM fine-tuning & evaluation, statistics/experimental design, clinical or healthcare NLP data experience, prompt engineering, data analysis & visualization

Key Skills

Education

Any Gradute