AI Security Fellow

Details

Anthropic’s mission is to create reliable, interpretable, and steerable AI systems. The Fellows Program accelerates AI security and safety research by providing funding and mentorship to promising technical talent for a four-month empirical research project. Fellows will primarily use external infrastructure (e.g. open-source models, public APIs) to work on an empirical project aligned with Anthropic's research priorities, with the goal of producing a public output (for example, a paper submission).

Responsibilities

• Conduct empirical AI security and safety research using external infrastructure (open-source models, public APIs).
• Work on a project aligned with Anthropic’s research priorities with the goal of producing a public output (paper, blog post, etc.).
• Collaborate with and receive mentorship from Anthropic researchers.
• Communicate research findings clearly; iterate quickly on experiments and implementations.

Requirements

• Fluent in Python programming (required).
• Available to work full-time on the Fellows program for 4 months (expectation: 40 hours/week).
• Strong technical background in computer science, mathematics, physics, cybersecurity, or related fields (or equivalent experience).
• Motivation to reduce catastrophic risks from advanced AI systems and interest in transitioning into empirical AI safety research.

Strong candidates may also have

• Contributions to open-source projects in LLM- or security-adjacent repositories.
• Experience with pentesting, vulnerability research, or other offensive security practices (e.g., reported CVEs, bug bounties).
• Experience with empirical ML research projects, deep learning frameworks, and experiment management.
• Demonstrated success tackling ambiguous technical problems and producing high-quality outputs.

Compensation

• Weekly stipend: 3,850 USD / 2,310 GBP / 4,300 CAN.
• Expectation of 40 hours per week.
• Fellows also receive funding for compute (~$15k/month) and other research expenses, plus access to benefits (benefits vary by country).

Logistics

• Program length: 4 months (with possible extension).
• Fellows will have access to a shared workspace in either Berkeley, California or London, United Kingdom. Anthropic is also open to remote fellows located in the UK, US, or Canada.
• To participate in the Fellows program, you must have work authorization in the US, UK, or Canada and be located in that country during the program. Anthropic is not currently able to sponsor visas for fellows.
• Applications and interviews are managed by Constellation (application link provided).
• Application deadline: Please apply by January 12, 2026.

Interview process

• Initial application and references check, technical assessments and interviews, and a research discussion.

Mentorship & Community

• Direct mentorship from Anthropic researchers.
• Potential mentors include Nicholas Carlini, Keri Warr, Evyatar Ben Asher, Keane Lucas, and Newton Cheng.
• Connection to the broader AI safety research community and a shared workspace where mentors may visit.

How to apply

• Apply via the Constellation application portal (link provided in the posting).