Anthropic Fellows Program, AI Safety & Security

Anthropic's Fellows Program is a 4-month, full-time research and engineering fellowship focused on AI safety and security. Fellows receive funding, mentorship, and access to a shared workspace in Berkeley or London, with remote options in the US, UK, or Canada. The program aims to produce public research outputs, with over 80% of prior fellows producing papers.

• Work on an empirical research project aligned with Anthropic's research priorities, using external infrastructure such as open-source models and public APIs.
• Collaborate with mentors on projects in areas like scalable oversight, adversarial robustness, model organisms, mechanistic interpretability, and AI welfare.
• Produce a public output, such as a paper submission.

• Strong technical background in computer science, mathematics, or physics.
• Fluent in Python programming.
• Available to work full-time for the program duration.
• Motivated by making AI safe and beneficial for society.
• Strong candidates may have experience in empirical ML research, working with large language models, or relevant disciplines like economics, social sciences, or cybersecurity.

• Weekly stipend of 3,850 USD / 2,310 GBP / 4,300 CAD, with expectation of 40 hours per week for 4 months.
• Funding for compute (~$15k/month) and other research expenses.
• Benefits vary by country.

Apply by October 30 via the provided link. The interview process includes an initial application and reference check, technical assessments and interviews, and a research discussion. Applications are managed by Constellation, Anthropic's recruiting partner.