Anthropic Fellows Program, ML Systems & Reinforcement Learning

Anthropic is recruiting for its Fellows Program, focusing on machine learning systems and reinforcement learning. Participants engage in full-time empirical research for a four-month period, utilizing external infrastructure such as open-source models and public APIs. The primary objective is to produce a public output, like a paper submission, aligned with the company’s research priorities. Fellows receive direct mentorship from Anthropic researchers and gain access to a shared workspace. The program also provides a weekly stipend, compute funding, and coverage for other research expenses. Applicants are not required to have previous experience, as the initiative aims to foster talent regardless of background. The selection process involves an initial application, reference checks, technical assessments, and a research discussion. This role is suitable for promising technical individuals interested in AI safety and security who wish to connect with a broader research community. The position is remote-friendly, with headquarters located in London, Ontario, San Francisco, and across the United States. Applications are reviewed on a rolling basis for a cohort expected to start in January 2027, though flexibility for different start dates may be possible.