Create reproducible reinforcement learning environments that test how AI models handle real software engineering tasks. This remote contractor role offers about 15 hours per week and pays $100 to $150 per hour.
The work
You will create structured reinforcement learning tasks that evaluate how AI models solve realistic software engineering problems. Your work will turn DevOps, CI/CD, debugging, performance, and command-line workflows into repeatable environments with clear expected behavior.
You will also create dependable reference solutions for evaluating model-generated approaches. The work may use Git, Docker, GDB, AddressSanitizer, FFmpeg, and other developer tools.
- Design software engineering workflows that test an AI model's ability to solve complex technical problems.
- Build reproducible environments with clear inputs, outputs, and expected behavior.
- Develop golden reference solutions for evaluating model-generated approaches.
- Work on DevOps, CI/CD, debugging, performance, and code-maintenance scenarios.
- Use sound engineering judgment to keep tasks maintainable, repeatable, and technically meaningful.
What it pays and takes
This is a part-time contractor engagement for an experienced software engineering contributor. Compensation terms should be confirmed before accepting the engagement.
- Pay: $100 to $150 per hour in USD.
- Schedule: Approximately 15 hours per week, or fewer than 20 hours.
- Location: Globally remote, with eligibility based on the countries listed for the role.
- Language: English.
- Programming: Proficiency in at least one of C++, Python, Java, Go, TypeScript, or Rust.
- Technical knowledge: Strong understanding of algorithms, data structures, and performance tuning.
- Experience: Ability to debug complex software issues and deliver maintainable solutions.
- Additional requirements: Experience with feature development, codebase refactoring, performance or scalability optimization, written and verbal communication, and collaborative software development.
- Helpful but not required: Experience with large-scale or distributed codebases, modern AI or machine learning systems, rigorous code reviews, and software engineering best practices.
How it works
Apply on OpenTrain with your resume, then complete the application on the hiring site.
About AI training work
AI training work uses human-designed examples and evaluations to help AI systems solve problems more reliably. OpenTrain is the hiring and contracting organization for this role, connecting contributors with projects where their technical judgment helps assess and improve AI systems.
