AI Safety Fund Grants
Mercor · San Francisco · Mexico | Mexico City · New York City · London
CHECKED 22H AGO
Applied AIonsite$0
At a glance
Highlights
- Fund frontier research on misalignment, scheming, and deceptive alignment.
- Build evals for sandbox escape, oversight, and evaluation awareness.
- Release open papers, datasets, and tools advancing AI safety verification.
What they offer
- Funding for researcher hours, API costs, and event and conference attendance
- Access to Mercor expert network for grading, red-teaming, and annotation
- Access to internal evaluation infrastructure and researcher network
Ideal candidate
A curious, mission-driven researcher passionate about making frontier AI safe and verifiable, eager to publish open work with real-world impact.
Requirements
- Researcher at university, nonprofit, public benefit corporation, or small startup not selling AI training data
- Based in the United States or the United Kingdom
- Background in ML, CS, statistics, or adjacent field
- Specific, well-scoped proposal with outputs, impact, and resource needs
- Willing to publish open paper, dataset, or tool under commercial-use license
- hiring process
- Submit a one- to two-page Expression of Interest covering team and background, proposed research project, and outputs with timelines.
Lab profile
About Mercor
AI data company organizing human expertise to train frontier models for the AI economy.
What they build
- APEX benchmark family
- Mercor Enterprise
- Human expertise training data platform
Known for
- Massive domain-expert network
- Frontier model training infrastructure
- Real-world professional work evaluation
Key facts
- Millions of domain experts on platform
- Pays over $4M per day to experts