My fork of the UofTHacks 2025 team project (team of 4). Canonical repo: tanujdargan/html.ai.
Real-time UI personalization: classifies a visitor's behavior and serves the best-performing content variant, adapting as they browse.
- Built the 2 demo e-commerce sites (Shoopify and Amazoon) that show the engine personalizing content across sites.
- Built the real-time analytics dashboard (user tracking, variant performance).
- Built the live debug panel that surfaces each visitor's variant changes in real time.
- Integrated the team's JS SDK (the
<ai-optimize>custom element) into both demo sites.
(The core FastAPI/Gemini backend was built by teammates; my work was the demos, dashboard, debug panel, and SDK integration.)
Python, FastAPI, Gemini 2.0 Flash, MongoDB, Docker, vanilla-JS SDK.
- Canonical team repo: https://github.com/tanujdargan/html.ai
- Live demo: https://htmlai.tech