Up to $167,000 Anthropic STEM Fellowship
- Organization: Anthropic
- Type: Fellowship
- Location: United States
- Deadline: October 28, 2026
- Funding: Fully Funded
Anthropic, the AI safety company behind Claude, runs a four-month fellowship that brings established scientists and engineers into the company to set the hardest open problems in their fields and test Claude against them. Fellows work in small pods by discipline, building datasets, simulations and evaluations that feed Claude's training. **What you get:** - $100,000 to $167,000 USD for the four-month fellowship - A place in a small pod of Fellows in related fields, meeting weekly with Anthropic researchers - Access to frontier Claude models and internal evaluation tooling **What you'll do:** - Define long-horizon open problems in your field where Claude falls short of expert judgment - Build datasets, simulations and hard-to-game evaluations that test whether models can plan experiments and interpret data - Apply Claude to open problems in your domain using the scientific tools your field relies on **Key dates:** - Applications open until October 28, 2026 - Offers planned by November 20, 2026 - One-week in-person orientation in San Francisco - Fellowship runs January 12 to May 8, 2027, full time **More information:** - Later stages typically include a take-home exercise and conversations with the research team - Default location is San Francisco; remote within the United States considered in certain circumstances - Fixed-term contract through an employer of record, not direct employment by Anthropic
