Technical Code Evaluation & RLHF for Front-End Frameworks
This project involves high-precision labeling, auditing, and Reinforcement Learning from Human Feedback (RLHF) for Large Language Models (LLMs) specializing in software engineering. My primary focus is on the evaluation of model-generated code snippets in HTML5, CSS3, and JavaScript (ES6+). I perform rigorous quality assessments to ensure generated code is syntactically correct, follows modern best practices (e.g., semantic HTML, responsive CSS), and is optimized for cross-browser compatibility. Beyond syntax, I provide expert-level feedback on logic flow, accessibility compliance (WCAG standards), and the integration of front-end libraries. By identifying edge cases and debugging erroneous model outputs, I help refine the model’s ability to generate production-ready code that adheres to professional industry standards.