You will play a key role in advancing AI systems that write production-quality code. Using your deep expertise in Java, you'll analyze, compare, and improve code snippets generated by AI models, ensuring they meet high standards for correctness, efficiency, and maintainability.
What You'll Do
- Evaluate multiple versions of AI-generated code, selecting the best option and clearly justifying your choice
- Refactor and fix code to resolve bugs, optimize performance, and align with best practices
- Provide structured feedback—ratings, edits, test outcomes—that feeds into reinforcement learning systems
- Write original code as part of the training process to guide model behavior
- Translate nuanced human judgment into actionable signals that shape model learning
- Help ensure the AI can effectively critique and iterate on its own output
What We're Looking For
- At least three years of professional Java development experience
- Strong ability to detect flaws in logic, performance, and security during code review
- Exceptional attention to detail and clarity in written explanations
- Comfort reading technical specifications and documentation
- Proven ability to work independently with minimal supervision
- Preference for asynchronous workflows and written communication
Nice to Have
- Experience with constraint programming
Work Environment
This is a fully remote position with no geographic restrictions. The role operates in a low-interruption, asynchronous setting where clear writing and self-direction are essential. You’ll contribute directly to model improvement with measurable impact—no bureaucracy, no filler tasks.
Compensation
Hourly rate ranges from $30 to $70, based on location and experience level. There is no equity or additional compensation components.
