EcoSnap — Recycling Web App, Winner in WiCS HopperHacks 2026 | React, Vite, Three.js, React Three Fiber, Gemini API
- Built an AI-powered web application that analyzes uploaded waste images with the Gemini API and provides real-time recycling and reuse recommendations.
- Developed an interactive gamification system where users earn points for recycling and grow a 3D virtual cherry blossom forest using Three.js and React Three Fiber.
- Implemented persistent user progress for recycling statistics, points, and forest growth while integrating the AI image-analysis workflow into a React/Vite frontend.
MoodMix — Reliability-Aware Music Recommender | Python, Pytest
- Developed a content-based recommendation engine that ranks songs using weighted genre, mood, and energy signals and generates explainable scoring for each result.
- Engineered a reliability gate and input-validation pipeline that limits or rejects low-confidence and invalid recommendations instead of returning arbitrary results.
- Built automated regression and end-to-end tests covering ranking, guardrails, and reliability behavior, achieving 9/9 unit checks and 6/6 evaluation cases.