About this role
Mercor is working with a leading AI research lab on a project to improve how AI systems do real-world software engineering work. We're looking for senior software engineers who have deep experience at top technology, quantitative or financial firms. You'll use that expertise to train and evaluate frontier AI models.
This role is open to candidates based in the United States, United Kingdom, Canada, Australia or New Zealand.
Key Responsibilities
- Designing realistic, challenging software engineering tasks drawn from real-world work, including but not limited to:
- feature changes in large codebases
- debugging and root-cause analysis
- code review
- system design and architecture decisions
- codebase Q&A
- technical documentation and planning
- Writing clear task proposals that explain why each task is realistic, how it is verified and what makes it hard for current AI models
- Illustrating "what good looks like" by writing reference solutions, tests and precise, task-specific criteria covering correctness, architecture, maintainability, style and scope
- Assessing AI-generated code and engineering work, and working alongside AI coding agents to judge where they succeed and fail
- Applying consistent, evidence-based judgment so that assessments are reproducible and defensible
- Taking structured feedback from senior reviewers and iterating on submitted work
Ideal Profile
- 4+ years of professional software engineering experience in one of these settings:
- A leading technology or quantitative firm (such as FAANG+, Magnificent 7+, top-tier quant funds or equivalent) with a Mid-senior level title. Machine Learning, Infrastructure, Data, Security and Quant Developer engineering roles may also qualify.