Anthropic Fellows Program, ML Systems & Reinforcement Learning

Overview

Anthropic Fellows Program workstream for ML Systems & Reinforcement Learning, offering 4 months of full-time AI research with mentorship and funding.

What you'll do

• Conduct an empirical project aligned with Anthropic research priorities.
• Use external infrastructure such as open-source models and public APIs.
• Work toward producing a public output (e.g., a paper submission).
• Select and match with a project mentor.
• Build and/or integrate ML systems components (e.g., simulators, accelerator backends, pipelines).
• Train, fine-tune, or evaluate large language models as part of research.

What you'll need

• Be fluent in Python programming.
• Have a strong technical background in computer science, mathematics, or physics.
• Be able to attend full-time for the Fellows program duration.
• Have strong software engineering skills for building complex ML systems.
• Have experience analyzing and debugging model training processes.
• Have experience with training, fine-tuning, or evaluating large language models.

Details

• Program expected to start in January 2027.
• Fellows work in a shared workspace in Berkeley or London, and remote fellows are accepted in the UK, US, or Canada.
• Work authorization required in the US, UK, or Canada, with location during the program.
• Visa sponsorship is not available for fellows.