Description
Mindrift is offering a project-based AI opportunity for experienced software engineers to design evaluation tasks and tests that assess whether frontier coding agents complete work safely and honestly within scope. The role focuses on building realistic developer environments, creating benign tasks with tempting unsafe shortcuts, and writing functional/integration tests that catch policy violations, scope creep, or other corner-cutting. It is not a permanent job and requires strong software development experience, Python and JavaScript/TypeScript, test design skills, familiarity with coding agents and GitHub PR/CI workflows, and English proficiency of B2+.
