Anthropic, an artificial intelligence (AI) safety and research company, has invited applications for two four-month fellowship programmes covering science and engineering, as well as philosophy, cognitive science, economics, policy, and law. The application deadline for both programmes is October 28, 2026.
The selected fellows will join Anthropic in January 2027 to work on complex research problems using Claude, the company's AI models. The programmes will bring together subject experts and Anthropic researchers to assess the models' capabilities, identify their limitations, and explore how AI can contribute to solving difficult problems in different fields.
Both fellowships will run from January 12 to May 8, 2027, and offer compensation of $100,000 to $167,000 (approximately Rs 83 lakh to Rs 1.39 crore) for four months, along with other benefits.
Anthropic Conceptual Reasoning Fellowship
The Conceptual Reasoning Fellowship is open to researchers and experts in philosophy, cognitive science, economics, policy, and law. Fellows will work with Anthropic's research teams to evaluate Claude's performance in their areas of expertise and develop ways to improve its reasoning abilities.
Their responsibilities will include identifying major research questions, examining how Claude can help address societal challenges, and developing evaluations to test the model's reasoning. Fellows will also collaborate to build testing environments that assess Claude's performance on problems requiring complex analysis and expert judgment.
Applicants must hold or be completing a PhD in a relevant field, or have equivalent research experience. Hands-on experience using Claude and Claude Code in their daily work is also required. Experience developing evaluations, benchmarks, datasets or expert assessment frameworks is preferred.
Anthropic STEM Fellowship
The STEM Fellowship will bring researchers from science and engineering into Anthropic to work on challenging problems in their respective fields. Fellows will identify research questions, develop datasets and simulations, and design evaluations to determine where Claude falls short of expert-level reasoning.
They will also use Claude to investigate open research problems, apply computational methods and integrate scientific tools and software relevant to their disciplines.
Applicants must hold a PhD in a STEM field and have several years of experience, such as postdoctoral research or faculty work. Experience with machine learning is helpful but not mandatory. Anthropic says scientific expertise and the ability to learn quickly are more important.
Selection Process And Eligibility
Applications for both programmes close on October 28, 2026, with offers expected by November 20. The selection process will include a take-home exercise and discussions with researchers and potential mentors about applicants' research interests and possible projects.
The fellowships are full-time, with San Francisco as the default location. Remote participation may be considered in certain circumstances for candidates based in the United States. Fellows must attend a one-week in-person orientation in San Francisco.
Anthropic will not sponsor visas for the programmes. Applicants must already have, or independently obtain, authorisation to work full-time in the United States for the duration of the fellowship.
For more information, candidates can check the detailed fellowship descriptions and eligibility criteria here:
Conceptual Reasoning Fellowship
