Software Engineer (Open Source Contributor)

Coding environments for AI evaluation

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

Work is remote worldwide and does not require previous AI experience. Strong open-source contributions and a profile demonstrating them, such as GitHub, help establish your technical background; a HackerRank assessment can be used when that evidence is unavailable.

What you will do

• Develop realistic, challenging coding tasks involving bug fixes and feature implementation.
• Contribute code samples, debugging strategies, and technical reasoning in Python3, Java, Rust, Go, C++, or TypeScript.
• Diagnose defects and performance bottlenecks across unfamiliar or complex codebases.
• Implement robust features with attention to scalability, maintainability, and software engineering practices.
• Refactor and optimize legacy code for clarity, efficiency, and reliability.
• Document solution approaches and code decisions so the resulting training data captures the reasoning behind the implementation.
• Review peer code and technical submissions for correctness and clarity.

Experience that fits

Strong backend experience may be in Java, Node.js, Go, Rust, TypeScript, Python, C++, or a similar language. The work calls for experience building scalable services and APIs, automated testing, debugging, performance work, and careful code review.

• Demonstrated delivery of reliable, performant backend projects and substantial features.
• Substantial expertise in at least one listed language.
• Strong understanding of algorithms, data structures, and core computer science concepts.
• Experience with complex bug resolution, large codebase refactoring, legacy modernization, and system optimization.
• Clear analytical, problem-solving, documentation, and technical communication skills.
• Interest in AI engineering challenges; prior AI work is not required.

Payment and timing

The advertised range is $50 to $100 per hour, with roughly 15 hours per run. Actual compensation is output based: contributors are paid for tasks that meet project specifications, and completion time varies by experience and workflow.

Minimum weekly task submissions are required, but the listing does not specify a count. Selected contributors receive onboarding materials and are expected to begin their first tasks within 24 to 48 hours after onboarding; immediate availability is preferred.

Selection

The process includes screening questions, an AI interview of approximately 20 minutes, and hiring manager review. A technical assessment may be required when a GitHub profile does not fully demonstrate the needed experience.

Interview and assessment work must be completed independently, without AI assistants or help from other people. Completing the role's AI interview can speed up the selection process.