jobs
Staff+ Software Engineer, Safeguards Evals
About Anthropic Anthropic’s mission is to create reliable, interpretable, and steerable AI systems. We want AI to be safe and beneficial for our users and for society as a whole.
Auf Englisch
- Company
- Anthropic
- Location
- New York City, NY; San Francisco, CA; Seattle, WA
- Status
- Open
- Posted
- 2026-11-09 20:05 UTC
Anthropic is hiring a staff-plus software engineer for safeguards evaluations. The role sits between applied machine-learning research and engineering: designing experiments on evaluation quality, building datasets that reflect real misuse, and creating harnesses and metrics for an agentic investigation system. The engineer would move successful methods into training, agent-change and release pipelines, and build tools for policy staff. The posting asks for data-pipeline experience, work with LLMs and agentic systems, strong analysis, and production coding. It prefers eight or more years of industry experience and evaluation or trust-and-safety background.
Original job posting