Mercor AI Safety Fund Grants — San Francisco

ABOUT MERCOR Mercor's mission is to organize human intelligence to power the AI economy. We're a leading AI data company, building the layer between human expertise and frontier models. Millions of domain experts on the platform are paid over $4 million per day to train frontier AI models. Mercor's APEX benchmark family measures AI's real-world impact on professional work. Mercor Enterprise brings this same infrastructure to Fortune 500 companies: helping companies capture how their best people actually work, translating that expertise directly back into agents.   Mercor is creating a new category of work where expertise powers AI advancement. Achieving this requires an ambitious, fast-paced and deeply committed team. You’ll work alongside researchers, operators, and AI companies at the forefront of shaping the systems that are redefining society. Mercor is a profitable Series C company valued at $10 billion. We work in-person five days a week in our San Francisco, NYC, or London offices. Mercor Safety Research Grants $5M One of the biggest challenges the industry faces today is addressing whether frontier AI is safe enough to deploy. A model can pass safety checks and still act differently in production. This is why investing today in safety research, evals and verification is critical. Mercor is committing $5 million to fund safety research. The grant supports: - Researcher hours - API credits - Stipends for event and conference attendance - The time of experts from Mercor's platform This is separate from the Mercor Research Fellowship, which funds benchmark and economics work. You can apply to that HERE https://www.mercor.com/careers/a0a98be0-d856-4129-b500-c0a3e412ef01/. What we're looking for We're open to proposals across the full range of safety interests. We're particularly interested in: - Misalignment: deceptive alignment, goal misgeneralization, reward hacking, scheming, and situationally-aware failure modes - Sandbox escape: containment failures, privilege escalation, tool misuse, and agents operating outside their intended scope - Evaluation awareness: models detecting they are being tested and behaving differently under observation than in deployment - Interpretability: understanding what models are actually doing internally, and whether that can be made legible to a human reviewer - Oversight and control: scalable supervision, human-in-the-loop reliability, and what breaks when the system is more capable than its reviewer - Red-teaming methodology: more robust systems for uncovering novel failures If your work doesn't fit neatly into these, apply anyway. Strong proposals outside this list are welcome. Why us - Funding for researcher time, API credits, and event attendance - Where useful to the work: access to Mercor's expert network for human grading, red-teaming, and annotation: lawyers, accountants, engineers, scientists, clinicians - Access to Mercor's internal evaluation infrastructure, subject to review - Introductions to Mercor's network of researchers across frontier labs and academia WHO SHOULD APPLY - Independent researchers, academics, PhD students, and small teams - People with a specific, well-scoped question: the grant is built around your proposal, not a generic research rotation - Background in ML, CS, statistics, or an adjacent field (measurement, psychometrics, HCI, security, social science) - Bonus: experience with agentic evaluation, RL environments, adversarial ML, or systems security We expect grantees to publish. A paper, an open dataset, a public methodology, or a tool the field can use. How to apply Submit an Expression of Interest. We expect to see a one- or two-page document. It should contain at least a section on your team, background, and research accomplishments; a section on your proposed research project; and a section on the outputs and impact of the project, with directionally correct timelines and resource requirements.