Projects

Senior thesis project: an AI study companion that recommends compatible study partners through a hybrid recommender over course data, stated preferences, behavioral signals, and schedule availability. Uses embeddings, similarity metrics, and constraint-aware scheduling to rank candidate matches, evaluated on precision, recall, and F1 against a profile-based baseline. Also adds Q&A, plain-language topic explanation, and study planning with reminders.

Full-stack capstone app built during Meta University: recommends personalized recipes from pantry contents, ingredient freshness, and user behavior to reduce food waste. Includes drag-and-drop meal scheduling, automatic grocery lists, timezone-aware expiration alerts, and adaptive recommendation/substitution algorithms. Reduced recommendation latency by 18% and improved recommendation accuracy by 24%.

Built with a 6-person team for the Alegeus challenge at the VISTA Hackathon: an AI chatbot answering customer questions about financial benefits, including COBRA enrollment. Integrated a conversational Chat API and designed JSON response handling so plan details returned as clear, structured answers rather than raw policy text.