Mercor Research Fellowship 2026

The Mercor Research Fellowship is a funded research programme for individuals to design and implement benchmarks, evaluation techniques, or empirical studies related to AI. Fellows work with the APEX research team, gain access to enterprise evaluation problems, and receive time, compute resources, mentorship, and potential long-term opportunities within Mercor. Focus areas include AI evaluation in professional services, software engineering, accounting and finance workflows, and AI-for-Science benchmarks among others.

Stipend (as above), API credits, compute budget, mentorship, access to data and networks, potential full-time offer consideration, and optional in-person desk at Mercor’s San Francisco office.

Applicants should have genuine interest in evaluation as a research discipline, a background in economics, CS, ML, statistics, or a related field, and a clearly defined idea for a benchmark, evaluation technique, or empirical study. ability to commit at least 30 hours per week for the duration.

Submit a well-scoped idea for a benchmark/evaluation technique or empirical study along with your background; follow the official apply link for submission.