Open Source Maintainer / Core Contributor – Software Engineer

Improve AI software engineering with repository experience

Contribute engineering and code review judgment to the development of foundational AI models. The work is intended for maintainers and core contributors who have substantial responsibility for a mature, widely used open source codebase.

This is full time W-2 employment through an employer of record partner, with a possible placement in a leading AI lab's extended workforce. The partner handles employment, payroll, benefits and compliance for the assignment.

Engineering and review work

• Help research and engineering teams address gaps in software engineering, code review and open source maintenance knowledge.
• Develop demanding but realistic features, refactors and bug fixes based on the repository you know best. Supply accurate reference solutions with clear structure.
• Assess code, tests and evaluation rubrics created by AI against the repository's design, established conventions and quality expectations.
• Build the project, execute tasks and run tests locally, using Docker where needed to verify the behaviour of solutions and tests.
• Explain what a successful implementation should accomplish and identify weaknesses in generated changes through precise written reasoning.
• Coordinate with other domain experts to keep training material consistent and technically correct.

Repository and contribution qualifications

You must qualify as a maintainer or core contributor to a widely used open source project. Maintainer evidence is a listing in MAINTAINERS or CODEOWNERS, or at least 10 pull request reviews during the previous 12 months; core contributors need at least three substantive merged pull requests over that period.

The repository must be at least two years old, have actual users and an active review process, and consist primarily of code written by people. Your own public GitHub history must demonstrate sustained contributions over multiple years.

Senior engineering judgment is needed to scrutinise generated code, test coverage and edge cases. You must be able to build and run both the project and its test suite locally, including with Docker, and communicate complex technical choices clearly. All rationales and feedback must be written personally and in your own words.

Engagement requirements

Commit to 40 hours each week on weekdays, with reliable, exclusive availability. During the engagement, you cannot hold another full time position or participate in other AI training or evaluation projects.

Any necessary background checks and confidentiality agreements must be completed before work starts. Work takes place on a device provided for the engagement to protect project information.