🛠 Tech Stack

💼 About This Role

You'll evaluate and refine code generation models by analyzing multi-step reasoning trajectories from real GitHub repos. Your work improves production-grade code outputs and model logic for complex programming tasks. This long-term project requires strong engineering judgment over surface-level labeling.

🎯 What You'll Do

• Review model-generated reasoning sequences for code
• Identify logical inconsistencies or weak reasoning steps
• Improve trajectory structure for production-grade outputs
• Evaluate reasoning quality across different programming environments

📋 Requirements

• Proficiency in at least 2 mainstream languages (Python, C++, Java, TypeScript, JavaScript)
• Real-world experience in backend systems, frontend applications, algorithms, testing, or infrastructure
• Comfortable reading and reasoning through large GitHub repositories
• Strong written communication skills

✨ Nice to Have

• Experience contributing to high-visibility GitHub repositories
• Proficiency in multiple programming paradigms

🚩 Heads Up

• Role mixes AI training with code generation, may be ambiguous
• Qualification exercise required before hiring