The work
You will evaluate language-model performance on real-world bug-fixing tasks in public GitHub repositories and open-source libraries. The work combines software setup, issue review, testing, and practical code execution.
You will also work with researchers to identify repositories and issues that create useful challenges for language models.
- Analyze and triage issues in open-source libraries.
- Configure repositories, including Docker-based development environments and local setups.
- Review unit-test coverage and overall test quality.
- Run and modify codebases locally to assess AI performance on bug-fixing tasks.
- Use Rust and related software engineering skills to understand complex projects.
What it pays and takes
This is a remote contractor assignment lasting approximately 2 to 3 weeks. Payment is made per approved task rather than as a standard hourly rate.
- $15 per approved task.
- 20+ hours per week.
- Remote work for contractors in India, Pakistan, Nigeria, Egypt, Ghana, Bangladesh, Turkey, and eligible countries in Latin America.
- At least 3 years of software engineering experience.
- Strong Rust experience.
- Fluency in English.
- Proficiency with Git, Docker, and basic software pipeline setup.
- Ability to navigate complex codebases and run, modify, and test real-world projects locally.
- Experience contributing to or evaluating open-source projects is helpful.
How it works
Apply on OpenTrain with your resume, then complete the application on the hiring site.
About AI training work
AI training uses human work to help artificial intelligence learn from examples, including software tasks and model evaluations. People with strong technical experience are needed to judge whether model outputs work correctly and to prepare reliable training data.