Week 2026-W02

Anthropic· 1 role

• Applied Safety Research Engineer, Safeguards Eval Gate · Research 9
  Research-oriented engineer to develop methods for representative, robust, and informative AI safety evaluations. This role involves designing experiments to improve model behavior evaluation, shipping these methods into pipelines that inform model training and deployment, and directly shaping how Anthropic understands and improves model safety across misuse, prompt injection, and user well-being. The role also involves building tooling for policy experts and surfacing findings to drive upstream model improvements.