Anthropic AI Safety Fellow, US

Details

Anthropic’s mission is to create reliable, interpretable, and steerable AI systems that are safe and beneficial for users and society. The team consists of researchers, engineers, policy experts, and business leaders dedicated to building beneficial AI.

Responsibilities

• Participate in the Anthropic Fellows Program to accelerate AI safety research through empirical projects using external infrastructure like open-source models and public APIs.
• Produce public outputs such as paper submissions aligned with research priorities.
• Receive mentorship from Anthropic researchers, funding, compute resources, and access to shared workspaces.
• Engage in a 2-month fellowship with possible extension to 6 months based on progress.

Mentors & Research Areas

Mentors include leading researchers working on topics such as:

• Scalable Oversight
• Adversarial Robustness and AI Control
• Model Organisms for misalignment
• Model Internals / Mechanistic Interpretability
• AI Welfare

Requirements

• Motivation to reduce catastrophic risks from advanced AI systems.
• Strong technical background in computer science, mathematics, physics, or related fields.
• Proficiency in programming, especially Python and machine learning frameworks.
• Availability to work full-time for at least 2 months (ideally 6 months).
• US, UK, or Canadian work authorization with ability to work either remotely (Canada) or out of Berkeley or London.
• Ability to work well in fast-paced, collaborative environments and independently execute projects incorporating feedback.

Strong Candidates May Have

• Experience with empirical ML research projects.
• Familiarity with large language models.
• Knowledge in interpretability and deep learning frameworks.
• Open-source contributions.

Interview Process

• Initial application and references by August 17.
• Coding assessment in Python.
• Coding-based technical interview.
• Final interviews including research discussion and take-home project.
• Offer decisions by early October with rolling basis.

Compensation

• Weekly stipend of $2,100 USD for 40 hours per week.

Location Policy

• Role can be done remotely anywhere in the US.
• Strong preference for Bay Area to access shared workspace.
• Expected at least 25% presence in offices for other staff, but exempt for this role.

Additional Information

• Bachelor's degree or equivalent experience required.
• Visa sponsorship offered selectively.
• Inclusive and diverse hiring encouraged.
• Focus on large-scale, collaborative AI research projects.
• Competitive compensation, benefits, flexible hours, and office collaboration space in San Francisco.