Artificial intelligence company Anthropic has opened applications for two research fellowship programmes for 2027, inviting experienced researchers in science, engineering, philosophy, cognitive science, economics, public policy and law to work with its AI models.
The STEM Fellowship and the Conceptual Reasoning Fellowship will run from January 12 to May 8, 2027. Both programmes will bring researchers into Anthropic's research teams to test the capabilities of its Claude AI models, identify weaknesses in their reasoning and develop ways to improve their performance on complex problems.
Selected fellows will receive compensation ranging from $100,000 to $167,000 for the four-month term, along with other benefits. Applications for both programmes close on October 28, 2026, with offers expected by November 20.
What are the two fellowships?
The two programmes target different areas of expertise.
The STEM Fellowship is intended for researchers in science and engineering who can use Claude to investigate difficult research questions. Fellows will work on problems in areas such as mathematics, physics, life sciences and engineering, developing datasets, simulations and evaluation methods to test whether AI models can reason through complex scientific problems.
Anthropic says previous machine-learning experience is helpful but not compulsory. Applicants must have a PhD in a STEM field and several years of relevant experience, such as postdoctoral research or faculty work.
The Conceptual Reasoning Fellowship is designed for experts in philosophy, cognitive science, economics, public policy and law. Fellows will work with Anthropic's researchers to assess Claude's reasoning in areas where answers often require specialist knowledge, interpretation and judgement.
Their work may include developing benchmarks, datasets and testing environments to measure how well the model handles complex questions. Applicants must hold or be completing a PhD in a relevant field, or have equivalent research experience.
How much will fellows be paid?
Both programmes offer compensation of $100,000 to $167,000 for the four-month fellowship, equivalent to approximately Rs 83 lakh to Rs 1.39 crore at recent exchange rates.
Anthropic lists the compensation as an annualised equivalent of $300,000 to $500,000, but the fellowship itself lasts four months. The advertised amount is for the fellowship period, not a full year's salary.
The positions are full-time, fixed-term contracts. Fellows will work primarily from Anthropic's San Francisco location, although remote arrangements within the United States may be considered in certain circumstances. All selected candidates must attend a one-week in-person orientation in San Francisco.
Who can apply?
The eligibility requirements differ between the two programmes.
For the STEM Fellowship, applicants must hold a PhD in a science, technology, engineering or mathematics discipline and have several years of research experience. For the Conceptual Reasoning Fellowship, applicants must hold or be completing a relevant PhD, or demonstrate equivalent research experience in the fields covered by the programme.
Candidates for both fellowships must have hands-on experience using Claude and Claude Code in their regular work. Experience designing evaluations, benchmarks, datasets or assessment frameworks is an advantage for the Conceptual Reasoning Fellowship.
Anthropic expects candidates to work on open-ended research problems, collaborate with its researchers and develop concrete ways to test the capabilities and limitations of AI models. The programmes are aimed at experienced researchers rather than applicants seeking a conventional entry-level internship.
Can applicants from India apply?
Applicants from India and other countries can apply if they meet the academic and professional requirements, but there is an important restriction on work authorisation.
Anthropic states that it will not sponsor visas for these fellowship positions. Candidates must already have, or independently obtain, authorisation to work full-time in the United States for the duration of the programme.
San Francisco is the default work location. Remote participation may be considered in certain circumstances, but only within the United States. Applicants based in India should therefore check whether they meet the work-authorisation requirements before applying.
How will the selection process work?
Applications close on October 28, 2026. Anthropic expects to make offers by November 20.
The later stages of selection will typically include a take-home exercise and discussions with the company's research team about the applicant's research interests and possible projects. Selected candidates will work alongside researchers and other fellows on problems relevant to their disciplines.
The fellowships form part of Anthropic's effort to bring specialist knowledge into AI research. By working with researchers who understand the problems in their respective fields, the company aims to develop more rigorous ways to test where its models perform well and where they still fall short.
Applicants can review the requirements and apply through the official listings for the STEM Fellowship and the Conceptual Reasoning Fellowship.