The work
You will create challenging software engineering tasks based on real production work, including feature changes, debugging, root-cause analysis, code review, system design, codebase questions, documentation and technical planning. Each proposal must explain why the task is realistic, how a solution can be verified and why current AI systems may struggle with it.
You will also produce reference solutions, tests and task-specific evaluation criteria covering correctness, architecture, maintainability, style and scope. Day to day, this includes reviewing AI-generated code, working alongside AI coding agents, documenting failure modes and revising submissions after structured reviewer feedback.
What Mercor screens for
The screening is likely to focus on verifiable senior engineering experience at a qualifying technology, quantitative or financial firm, plus specific examples from large and complex production codebases. Candidates should be prepared to defend technical decisions under follow-up and demonstrate strong command of at least one major language, such as Python, Java, C++, Go, TypeScript or Rust. Evaluation judgment matters as much as implementation skill, particularly the ability to distinguish passing tests from sound architecture, maintainability and correct scope.
Logistics
Candidates must be based in the United States, United Kingdom, Canada, Australia or New Zealand. The listing does not disclose whether the engagement is remote, whether work is completed asynchronously or the typical weekly hours, so candidates should confirm these points during screening and state their availability precisely. Pay band: $120–180/hr
