Remote job
Principal Harness Engineer
Job details
About this role
Role overview Build the engineering systems that help AI coding agents produce dependable, maintainable software. The focus is on creating reusable guidance, automated checks, evaluation methods, and release controls so teams can verify agent-generated changes efficiently and route higher-risk decisions to people.
Responsibilities - Design, deploy, and support shared harness components, including agent guidance, architectural rules, reference materials, and code transformations. - Create automated feedback such as linters, static analysis, architecture tests, verification loops, and model-assisted reviews. - Define quality gates, release criteria, and escalation boundaries for agent-produced changes. - Develop testing and evaluation infrastructure to check generated work against quality and safety expectations. - Improve repository documentation and context, and turn repeated agent mistakes into lasting controls. - Measure agent-assisted delivery and quality, then use the results to guide engineering investments.
Requirements - Degree in computer science or a related field, or equivalent commercial experience in software architecture, development, and testing. - Deep software engineering experience with large codebases and a strong grounding in architecture, testing, and maintainability. - Hands-on experience with AI coding agents and an understanding of their strengths and limitations. - Ability to build engineering tools using linters, static analysis, CI/CD, containerized build and test environments, and observability. - Experience with specification-driven development, context engineering, agent orchestration, engineering fitness functions, or developer platforms. - Systems-oriented judgment: ability to define quality as repeatable, enforceable rules and choose suitable controls.
Nice to have - Familiarity with agent instruction conventions such as AGENTS.md and contract testing tools such as Pact. - Experience mentoring engineers in harness practices and responsible use of AI tools.