We are looking for experienced software engineers to help train and evaluate AI models on real-world programming work. You will build and validate coding tasks, review AI-generated code, and write the reference solutions and tests that models are measured against.
About the role
We are looking for experienced software engineers to help train and evaluate AI models on real-world programming work. You will build and validate coding tasks, review AI-generated code, and write the reference solutions and tests that models are measured against. Your engineering judgment directly shapes how reliably AI systems write, debug and refactor code.
Key Responsibilities
Design realistic coding tasks (bug fixes, feature implementation, refactoring, performance optimisation) with clear specifications.
Build reproducible environments (Git, Docker) and write automated test suites that verify solutions.
Write high-quality reference solutions and explain the design decisions behind them.
Evaluate AI-generated code for correctness, performance, security and readability.
Rank and compare model outputs and justify your judgments in clear written rationale.
Identify recurring failure patterns in model output and report them to the project team.
Follow project guidelines, take part in calibration and review sessions, and incorporate feedback to keep quality consistent.
Official source: www.rex.zone. Details can change; the platform's page is the final word. How we check listings.
