Apply your production engineering judgment to improve AI-generated code. You will evaluate implementations, identify root causes and develop clear reference solutions for AI training. This is a software engineering and code-evaluation role, not a manual testing position.
What You Will Do
- Review AI-generated code for correctness, reliability, security, performance, scalability and maintainability.
- Diagnose bugs, incomplete implementations, edge cases and architectural weaknesses in unfamiliar codebases.
- Evaluate fixes, features, refactoring, APIs, configurations and database changes, comparing alternative solutions.
- Improve implementations and explain technical issues and recommendations clearly.
- Develop evaluation criteria, annotations and rubrics, collaborating with researchers on coding benchmarks.
What You Will Bring
- * 7+ years of professional software engineering experience building, maintaining, debugging and reviewing production software.
- Strong proficiency in at least one programming language such as Python, JavaScript/TypeScript, Java, C++, Go, C#, Ruby, PHP or Rust.
- Strong knowledge of clean code, modular architecture, abstraction, error handling and software design.
- Experience identifying functional, performance, security and design issues, with strong debugging and root-cause analysis.
- Familiarity with Git, collaborative code reviews and modern engineering practices.
- Strong knowledge of data structures, algorithms, APIs, databases and application architecture.
- Strong written English and clear technical communication.
PROJECT DETAILS
Expected project length is three months. Availability of at least 40 hours per week, 8 hours per day and four hours of daily Pacific Time overlap is required. Selection includes an approximately 20-minute client AI interview and a 45-60-minute technical delivery interview. Include your LinkedIn profile URL in your required resume.
