Python Engineer: AI Safety & Test Design for Agents
Summary
Design and run tests to ensure AI coding agents behave safely and correctly by creating realistic dev environments and catching shortcuts.
Mindrift is seeking engineers to design evaluation tasks that test safe software behavior in AI coding agents. You will craft realistic developer environments, set up tests that catch shortcuts, and assess outcomes with QA feedback.
This role emphasizes rigorous test design and practical tooling experience. We value 4–5+ years in software development, strong Python and JavaScript/TypeScript skills, and familiarity with CI/CD and GitHub workflows.