Medical AI Evaluation Physician Subject-Matter Expert
Document
RLHFServed as a board-certified physician subject-matter expert improving large language model performance on clinical reasoning and medical question-answering tasks. Work included preference ranking of model-generated responses, authorship and validation of evaluation rubrics, and structured assessment of clinical accuracy, safety, and guideline-concordance. Developed reference-standard answers and anchor cases used to calibrate model outputs, and contributed to red-teaming and factuality review for health-information reliability. Emphasis throughout on clinical soundness, appropriate handling of uncertainty, and alignment with current standards of care across multiple medical specialties.
2024 - Present