The work

You will turn realistic backend engineering problems into reproducible reinforcement learning environments. Day to day, that means designing feature and bug-fix tasks, preparing a golden reference solution, building deterministic verifiers, investigating defects and bottlenecks, refactoring legacy code, and reviewing submissions for correctness and clarity.

Technical profile

The platform looks for evidence of backend engineering, scalable services and APIs, debugging, performance analysis, automated testing and sound code review judgment. An open source profile is preferred when it showcases C++, Python, Java, Go, TypeScript, or Rust; candidates without sufficient public evidence may complete a HackerRank assessment. Strong backend experience is sought in Java, Node.js, Go, Rust, Typescript, Python, C++ or similar. Task contributions cover Python3, Java, Rust, Go, C++, or TypeScript, with significant hands-on expertise in at least one of those languages preferred.

Screening

The process includes application questions, an AI-led interview of approximately 20 minutes, a possible technical assessment and hiring manager review. Expect follow-ups on specific contributions, debugging decisions, verifier design and how you distinguish a valid alternative solution from one that only passes superficial tests. The interview and assessment must be completed without AI assistants or help from other people.

Logistics

This is a remote contractor role open worldwide. Work is task-based and largely asynchronous, with each run estimated at about 15 hours; minimum weekly submission requirements apply, although the listing does not specify the number. Selected contributors are expected to begin their first tasks within 24–48 hours after onboarding, and compensation is output-based for tasks that meet project specifications.

Pay band: $50–100/hr