Staff Software Engineer, Safeguards Evaluations
AnthropicNew York, NY / San Francisco, CA / Seattle, WA
AI safety & policy
Full-time
AI safety & policy
Full-time
Salary
$320,000 $485,000 USD per year
Routes to impact
Direct high impact on an important cause
Description
Build evaluation methods, datasets, and infrastructure for AI safety assessments and agentic investigation systems.
- Design experiments to improve evaluation quality, grading accuracy, and coverage across misuse, prompt injection, and well-being
- Build and own eval harnesses for long-horizon agentic investigation systems, including metrics, test cases, and RL environments
- Construct datasets grounded in real-world misuse, including cyber attacks, bio weapons, and influence operations
- Ship evaluation methods into training, regression, and release pipelines, and build tooling for policy experts
- Visa sponsorship available, with reasonable effort to support successful candidates
This text was generated by AI. If you notice any inconsistencies, please let us know using this form.
