The work
You will help build and check training datasets based on real software engineering problems. The work uses public GitHub repositories and their histories to create tasks that can be verified by human reviewers.
You will also assess how well language models understand code and fix bugs. You may work with researchers on task design and coordinate junior engineers when a project needs extra support.
- Analyze and triage issues in trending open-source libraries.
- Set up repositories and local development environments, including Docker and basic software pipelines.
- Review unit-test coverage and judge whether tests are useful and reliable.
- Change and run codebases locally to evaluate language model bug-fixing performance.
- Find repositories and issues that provide meaningful challenges for language models.
- Help researchers design software engineering tasks and datasets.
- Lead or coordinate junior engineers when project needs require it.
What it pays and takes
This is a fully remote individual contractor assignment. The listing does not specify a pay rate.
The role is listed as entry level, but the requirements call for at least three years of professional software engineering experience. English is the working language, and applicants must be based in one of the listed countries.
- Pay: Not specified in the listing.
- Time: At least 20 hours per week, with options for 20, 30, or 40 hours.
- Schedule: Four hours of daily overlap with Pacific Time.
- Location: India, Pakistan, Nigeria, Kenya, Egypt, Ghana, Bangladesh, Türkiye, or Mexico.
- Language: English.
- Required: Strong C# experience; proficiency with Git and Docker; and familiarity with basic software pipeline setup.
- Required: Ability to navigate complex codebases and run real-world projects locally.
- Required: Good judgment about GitHub issues, repository quality, test coverage, and expected software behavior.
- Helpful: Experience contributing to or evaluating open-source projects.
- Helpful: Previous LLM research or evaluation work, or experience building or testing developer tools or automation agents.
How it works
Apply on OpenTrain with your resume, then complete the application on the hiring site.
About AI training work
AI training work is the human side of building artificial intelligence. People review examples, code, and model outputs so systems can learn what useful and correct behavior looks like; this role focuses on software engineering tasks and bug-fixing evaluation.