Software Engineer | Remote - Contract

Xperteez Technology • ,

1d • $50 - $100 • Remote

About The Position

As a contractor, you will create reinforcement learning environments that assess AI models’ ability to solve real software engineering problems, including finding subtle bugs, building non-trivial features, refactoring code, and optimizing performance. Each task involves defining a problem, building a reproducible environment, and writing a reference solution.

The work is project-based and paid per approved task. The stated commitment is approximately 15 hours per week. Trainers are allocated five hours per task and reviewers 1.5 hours per task. The role is fully remote.

Experts create realistic, challenging coding tasks based on repositories they know well, build reproducible environments and deterministic verifiers, provide reference solutions, and flag tasks that are ambiguous or unverifiable. The selection process includes an application and screening questions, an AI interview, profile review, and onboarding. Selected candidates are expected to begin their first tasks within 24–48 hours of completing onboarding.

Requirements

• Hands-on expertise in at least one of Python, Java, Rust, Go, C++, or TypeScript.
• Skills in Python3, Java, Rust, basic C++, and TypeScript are listed.
• Knowledge of algorithms.
• Ability to fix bugs, implement features, refactor codebases, and optimize performance.

Nice To Haves

• Deep understanding of algorithms, data structures, and software engineering principles.
• Proven ability to debug complex systems, resolve bugs, and deliver effective optimizations.
• Experience with large codebase refactoring and legacy system modernization.
• Track record of delivering high-impact features from conception through delivery.
• Strong documentation and communication skills to articulate technical concepts clearly.
• Interest in AI and its technical challenges; previous AI experience is not required.

Responsibilities

• Create realistic, challenging coding tasks for AI models based on repositories they know well.
• Design tasks involving subtle bugs, non-trivial features, or meaningful performance work rather than lookup or boilerplate.
• Build a reproducible environment for each task.
• Write deterministic verifiers that accept correct solutions across valid implementations and reject incorrect, incomplete, or hard-coded approaches.
• Provide a reference solution for each task.
• Flag tasks that are ambiguous or unverifiable rather than submitting them.

What This Job Offers

• Education Level: No Education Listed