Clinical AI Research Fellowship

Role overview Help assess and improve AI systems for clinical and healthcare tasks as part of a part-time research fellowship. Working alongside researchers, engineers, and clinicians, you will develop realistic evaluation tasks, examine model performance and failure modes, and contribute to research on safer, more reliable healthcare AI.

Responsibilities - Create tasks and benchmarks grounded in real clinical workflows. - Evaluate advanced AI models on complex clinical tasks and analyze where and how they fail. - Design end-to-end experiments and suggest research directions for assessing model capabilities. - Work with AI researchers to refine evaluations and research contributions. - Attend weekly check-ins and present work in group sessions.

Requirements - Clinical background as a fourth-year medical student, PGY-2 or later resident, or a physician, nurse practitioner, physician assistant, registered nurse, or PharmD. - Experience working with sensitive healthcare data.

Nice to have Familiarity with reinforcement learning environments, model evaluations, or evaluation rubrics is useful but not required.

Benefits and work setup Part-time for 3–6 months, with a minimum commitment of 10 hours per week. The position is in person in San Francisco or remote within the U.S. The fellowship offers a stipend of $1,500 per week, API credits, GPU support, human data budgets, mentorship, and evaluation infrastructure. Fellows may learn from AI researchers, co-author peer-reviewed publications, and receive invitations to research events.