AI Training Data Creator — Hint Writing & Model Guidance
In this project, I am directly involved in generating detailed, diverse, and educational hints that help train AI models to solve complex reasoning and math problems. Each task requires careful understanding of the prompt, step-by-step guidance writing, and iterative testing to ensure the model reaches the correct answer without explicitly providing it. My responsibilities include: Analyzing mathematical and logical prompts, often involving functions, roots, and symmetry. Writing clear, non-trivial hints that progressively lead the AI to the known final answer. Ensuring hints cover multiple reasoning aspects and diverse explanatory approaches. Reviewing the model’s responses based on my hints and refining them if needed. Evaluating whether the ground truth answers are correct and marking discrepancies. This project focuses on enhancing AI reasoning and problem-solving capabilities by producing high-quality training data for supervised fine-tuning and reinforcement learning from