Anthropic Fellows Program, AI Safety

About the role#

The Anthropic Fellows Program is a 4-month, full-time research initiative designed to foster technical talent in the field of AI safety. Fellows gain direct mentorship from Anthropic researchers and work on empirical projects aligned with the company's research priorities. The program aims to produce public outputs, such as research paper submissions, and connects participants to the broader AI safety and security research community.

What you'll do#

• Conduct 4 months of full-time research.
• Participate in a project selection and mentor matching process.
• Execute an empirical project focused on areas such as scalable oversight, adversarial robustness, model internals, or AI welfare.
• Produce a public research output, such as a paper submission.
• Contribute to a collaborative research environment.

What you'll need#

• A strong technical background in Computer Science, Mathematics, Statistics, Physics, or Cybersecurity.
• Fluency in Python programming.
• A strong motivation to ensure AI systems are safe and beneficial for society.
• The ability to implement ideas quickly and communicate research findings clearly.
• An interest in transitioning into empirical AI research.

Location & details#

• Term: Fall 2026.
• Work Modality: On-site.
• Locations: California, United States and Ontario, Canada.
• Employment: This is a full-time, paid position.
• Note: Visa sponsorship is not provided for this program; candidates must possess independent work authorization in the United States or Canada.