Anthropic opens paid AI research residency to senior technical leaders

Experienced professors, founders and engineers now have a paid route into AI research without first becoming machine-learning specialists. Anthropic opened applications on November 2 for its Anthropic Residency, offering fixed-term research placements to established technical leaders who want to change fields. Applications close November 22, 2026, and the program is intended to support a full-time career transition.

A career change, not an entry-level program

The target applicants include professors in quantitative disciplines, technical founders and staff-level engineers. Anthropic wants people who have already led ambitious technical work and can implement ideas quickly, communicate clearly and collaborate in a fast-paced environment. A PhD, published papers and prior machine-learning experience are not required.

Anthropic’s rationale is that senior researchers from other fields bring skills that matter in AI: choosing worthwhile problems and shaping a research agenda. The company argues that these abilities are becoming more important as models improve. It points to researchers, including its own, who entered machine learning from physics, mathematics, neuroscience and software engineering.

This round adapts the model of the Anthropic Fellows Program for established technical leaders. Anthropic cites past fellows’ contributions to subliminal learning, the Petri auditing tool and persona vectors as examples of newcomers contributing to AI research. That is the company’s stated basis for the approach, rather than a measured outcome for this Residency round.

Research support alongside the field change

Residents will work on empirical research—questions explored through experiments—in collaboration with Anthropic’s research teams. Researchers will outline problems and match them to each resident’s interests and background. Each resident will have an Anthropic mentor, regular one-on-one guidance, access to frontier models and a dedicated computing budget for their project.

The program also sets aside time for self-study, talks and guidance on modern machine learning and AI safety. Potential mentors include Alignment Science lead Ethan Perez, Scalable Oversight lead Samuel Marks and AI security researcher Nicholas Carlini. The list names possible mentors; individual pairings are not presented as guaranteed.

The published research areas span safety, security and the systems used to train advanced models:

• AI safety and alignment: keeping models helpful and honest, understanding their internal workings, studying possible failures, building safeguards and investigating potential AI welfare.
• AI security and adversarial testing: stress-testing frontier models to understand their capabilities in areas including cybersecurity, robotics and cryptography.
• Machine-learning systems and reinforcement learning: improving model training and its infrastructure, including model architecture, training data, hardware backends and environments used for learning.

Flexible placements, with a San Francisco preference

Applicants can state a preferred placement length, and Anthropic says it will aim to arrange flexible start dates. The company prefers residents to work from its San Francisco offices, but remote participation is possible, especially for shorter placements. That flexibility does not extend automatically to a later job: most full-time Anthropic research roles require relocation to San Francisco.

Anthropic’s stated goal is for every resident to finish in a position to continue AI research full-time, either at the company or elsewhere. The intended destination is therefore broader than an Anthropic role. Applicants should be motivated by making AI safe and beneficial for society and interested in moving into empirical AI research full-time.

The announcement provides separate US and UK application postings. Anthropic particularly encourages applications from groups underrepresented in technology and says candidates may also be considered for its Fellows Program.