MATS is a 12-week fully-funded research fellowship for aspiring AI alignment, control, security, and governance researchers. Fellows work directly with mentors from Anthropic, Google DeepMind, OpenAI and other top organizations—with housing and office space provided in Berkeley or London. MATS has accelerated 450+ researchers; among alumni who graduated before 2025, 80% now work on AI safety.

Website: https://www.matsprogram.org/apply?utm_source=cstheory-jobs&utm_medium=job-board&utm_campaign=s26
Email: operations@matsprogram.org