Anthropic has opened applications for two four-month fellowships spanning STEM, philosophy, economics, policy and law, offering selected researchers up to $167,000 in compensation.
Artificial intelligence company Anthropic has invited applications for two research fellowship programmes designed to bring subject-matter experts together with its AI researchers. The programmes will focus on science and engineering, as well as philosophy, cognitive science, economics, policy and law.
Both fellowships will run from January 12 to May 8, 2027. Selected participants will work with Claude, Anthropic’s AI models, to explore complex research questions, assess AI capabilities and identify areas where further improvements are needed.
The application deadline for both programmes is October 28, 2026. Fellows will receive compensation ranging from $100,000 to $167,000 for the four-month period, equivalent to approximately ₹83 lakh to ₹1.39 crore, along with additional benefits.
The Conceptual Reasoning Fellowship is intended for researchers and professionals specialising in philosophy, cognitive science, economics, public policy and law.
Participants will collaborate with Anthropic’s research teams to examine how effectively Claude handles complex questions within their areas of expertise. Their work will help identify the model’s strengths, limitations and opportunities for improving its reasoning abilities.
Fellows will identify important research questions and investigate whether AI can help address challenging problems across their disciplines. They will also develop evaluation methods to measure Claude’s performance on tasks that require specialised knowledge and critical thinking.
Additional responsibilities will include:
Designing tests to assess Claude’s reasoning and analytical capabilities.
Building evaluation environments for problems requiring expert judgment.
Examining how AI could contribute to addressing broader societal challenges.
Working with Anthropic researchers to improve model assessment methods.
Applicants must hold, or be completing, a PhD in a relevant discipline or demonstrate equivalent research experience.
Practical experience using Claude and Claude Code in regular professional or research activities is also required. Candidates with experience developing benchmarks, datasets, evaluation tools or expert assessment frameworks may have an advantage.
The STEM Fellowship is designed for researchers working in science and engineering who want to investigate difficult technical questions with the support of Anthropic’s AI models.
Selected fellows will work alongside the company’s research teams to explore scientific problems, test Claude’s performance and determine where its capabilities fall short of expert-level reasoning.
Participants will identify promising research questions and develop the tools needed to investigate them. This may involve creating datasets, building simulations and designing evaluations to assess the model’s performance in specialised scientific and engineering tasks.
Fellows will also use Claude to explore unresolved research problems, apply computational techniques and integrate relevant scientific software and tools into their workflows.
The programme aims to assess how AI can contribute to technical research while identifying limitations that require further development.
Applicants must hold a PhD in a science, technology, engineering or mathematics discipline and have several years of relevant experience, such as postdoctoral research or university faculty work.
Prior machine-learning experience may be beneficial but is not compulsory. Anthropic places greater emphasis on strong scientific expertise, research ability and the capacity to learn new methods quickly.
Both programmes offer competitive compensation for a four-month commitment.
|
Fellowship detail |
Information |
|---|---|
|
Duration |
January 12–May 8, 2027 |
|
Compensation |
$100,000–$167,000 |
|
Approximate Indian currency |
₹83 lakh–₹1.39 crore |
|
Application deadline |
October 28, 2026 |
|
Expected offers |
By November 20, 2026 |
|
Start date |
January 12, 2027 |
The compensation figures cover the four-month fellowship period. Additional benefits are also available, although specific details may vary according to the programme’s terms.
Applications for both fellowships must be submitted by October 28, 2026. Anthropic expects to issue offers by November 20, giving successful candidates time to prepare for the January start date.
The selection process will include a take-home exercise, followed by discussions with researchers and potential mentors. These conversations will focus on applicants’ research interests, relevant expertise and possible projects they could pursue during the programme.
Candidates should be prepared to explain how their experience aligns with the fellowship’s objectives and how they would approach complex research questions using AI tools.
The fellowships are full-time programmes, with San Francisco designated as the default location. Remote participation may be considered in certain circumstances for candidates based in the United States.
Regardless of the working arrangement, participants must attend a one-week in-person orientation in San Francisco.
Anthropic will not sponsor visas for these programmes. Applicants must already possess, or independently secure, authorisation to work full-time in the United States throughout the fellowship period.
Candidates should therefore review the location and work-authorisation requirements before applying.
Anthropic’s fellowship programmes aim to bring specialised academic and professional knowledge into AI research. By working directly with experts across scientific, technical, legal and social disciplines, the company intends to evaluate how Claude performs on demanding real-world problems.
The programmes also provide researchers with an opportunity to use advanced AI models in their fields while helping identify weaknesses in reasoning, analysis and problem-solving.
Conclusion
Anthropic’s two 2027 fellowships offer researchers an opportunity to collaborate with AI specialists on complex problems across STEM, philosophy, economics, policy and law. With compensation of up to $167,000, the programmes combine interdisciplinary research with AI evaluation. Interested candidates must apply by October 28, 2026, and meet the relevant academic, experience and US work-authorisation requirements.