1. Design and refine LLM prompts to identify math knowledge gaps through student contest performance and diagnostic assessment analysis. 2. Track prompt versions and evaluate output quality to improve AI-driven learning insights. 3. Develop a question recommendation engine that maps identified learning gaps to the question database. 4. Recommend questions based on appropriate difficulty levels, topics, and student grade requirements. 5. Define and monitor success metrics for AI features, including recommendation relevance, student improvement rates, and model accuracy. 6. Analyze AI performance outcomes and document findings for effective reporting and handoff processes. 7. Support the optimization of AI-driven educational workflows and personalized learning experiences.