Anthropic Fellows Program for AI Safety Research (May & July 2026 cohorts)

About this program

Anthropic is offering the Anthropic Fellows program, a four-month funded mentorship track (cohorts beginning May and July 2026) for engineers and researchers to pursue empirical AI safety projects aligned with Anthropic’s research priorities. Fellows receive a weekly stipend (US$3,850 / £2,310 / CA$4,300), roughly US$15k/month in compute funding, and close mentorship, with the program geared toward producing public outputs such as papers. Applicants should demonstrate strong technical fundamentals (comfortable coding in Python), the ability to turn ambiguous problems into concrete progress, and a clear motivation to reduce catastrophic AI risk. Formal credentials are not required—PhDs or prior ML publications are unnecessary—and candidates from diverse quantitative backgrounds are encouraged to apply.