Mercor Research Fellowship in London 2026 | Call for Applications

Mercor is accepting applications for its Research Fellowship program. This fellowship funds researchers building next-generation AI evaluation benchmarks. Fellows also conduct economic analyses examining AI’s labor market impact. Mercor operates APEX, a benchmark family testing real professional AI capabilities. This includes evaluating multi-hour agentic tasks across law and finance. The company recently reached a $10 billion valuation as a Series C startup.

Program Snapshot

• Host Organization: Mercor
• Locations: San Francisco, London, or New York City, remote allowed
• Duration: Three to six months, rolling admission
• Stipend: $40,000 for three months, or $80,000 for six months
• Application Deadline: September 26, 2026, 6:00 AM GMT

Focus Areas

• Long-horizon agentic tasks in professional services
• Real-world software engineering evaluation methods
• AI-for-Science benchmarks in math and biology
• Novel evaluation methodology and economic research

About the Fellowship

This program is built around individual research pitches, not generic rotations. Applicants propose a specific benchmark, evaluation technique, or economic study. If selected, fellows receive compute resources, expert labor, and dedicated mentorship. This supports designing, building, and publicly releasing their proposed research fully.

Fellows work directly alongside Mercor’s experienced APEX research team throughout. They also gain access to real evaluation problems from Fortune 500 partners.

What Fellows Will Do

Selected fellows design task specifications and grading rubrics with domain experts. This includes lawyers, accountants, engineers, and scientists across various fields. Fellows then pilot tasks, calibrate scoring, and test frontier models directly. On the economics side, fellows might analyze Mercor’s proprietary marketplace data. Final outputs may include published papers, open datasets, or new leaderboards.

Who Should Apply

Ideal candidates hold genuine interest in evaluation as a research discipline. Relevant backgrounds include economics, computer science, statistics, or social science. Applicants need a well-scoped, specific idea rather than a broad direction. Comfort with fast-paced, less-structured startup environments matters significantly here. Applicants must commit at least thirty hours weekly throughout their fellowship term.

Compensation and Benefits

Fellows receive unlimited API credits alongside dedicated GPU compute budgets. Weekly one-on-one mentorship comes from an APEX research team member directly. Fellows also gain access to frontier model APIs and evaluation infrastructure. Standout fellows may receive a full-time offer following their fellowship term.

How to Apply

Submit your specific research pitch through Mercor’s official careers portal.

#339966;">#339966;" href="https://www.mercor.com/careers/a0a98be0-d856-4129-b500-c0a3e412ef01/" target="_blank" rel="noopener">Click Here to Apply

Conclusion

This fellowship offers exceptional access to real enterprise AI evaluation challenges, backed by substantial funding and direct mentorship from an industry-leading research team. Its emphasis on funding specific, well-scoped ideas rather than generic research makes it especially suited to applicants with a clear, original proposal already in mind.

Given the pitch-driven selection process, prospective applicants should focus on articulating one sharp, well-defined research idea before the September 26 deadline.

#008000;">READ MORE