Software Engineer (GitHub Contributor)

Coding environment development

Build reproducible reinforcement learning environments that test whether an AI model can fix bugs, add features, refactor code and improve performance. Each problem needs a reference solution and deterministic checks that accept reasonable valid solutions while rejecting incorrect ones.

This contractor role is remote worldwide. Prior AI experience is not required, but practical backend judgment, debugging skill and clear technical explanations are central to the work.

Responsibilities

• Create realistic, challenging coding tasks involving feature implementation, bug fixes and performance improvements.
• Contribute code samples and explain development and debugging strategies in Python 3, Java, Rust, Go, C++ or TypeScript.
• Investigate complex defects and performance bottlenecks across codebases.
• Implement new features with attention to scalability, maintainability and software quality.
• Refactor and optimize legacy code for clarity, efficiency and reliability.
• Document the reasoning behind solutions and code decisions so the resulting training data is understandable and useful.
• Review peer code and technical submissions for accuracy and clarity.

Engineering background

Strong backend experience is expected in languages such as Java, Node.js, Go, Rust, TypeScript, Python or C++. Relevant experience includes scalable services and APIs, automated tests, debugging, performance analysis and thoughtful code review.

Preferred qualifications include a strong grasp of algorithms and data structures, work on substantial codebases, complex feature delivery and legacy system modernization. You should be able to explain technical decisions clearly and show a track record of delivering reliable, efficient systems.

Evidence of experience and selection

Clear open source contributions and a profile showcasing them, such as GitHub, are requested. Contributions in C++, Python, Java, Go, TypeScript or Rust are preferred; candidates without suitable evidence can instead be assessed through HackerRank.

Selection includes screening questions, an approximately 20 minute AI interview, a technical assessment where the profile does not fully demonstrate the needed experience, and hiring manager review. Completing the role's AI interview can accelerate the process. Interviews and assessments must be completed independently, without AI assistants or help from others.

Workload, payment and start

The listing gives a rate of $50 to $100 per hour and an estimated 15 hours per run. Actual compensation is based on tasks that meet project specifications, with completion time varying by experience and workflow.

Minimum submission requirements apply, including a minimum number of tasks each week.

Onboarding information is provided after selection. Roles are typically filled within 48 hours, and selected contributors are expected to begin their first tasks within 24 to 48 hours after onboarding.