Architecture

How this site is built, thinks, and plays, in diagrams.

This site is four systems wearing one documentation skin: a set of static pages, a retrieval-grounded chatbot, a reinforcement-learning game, and a self-hosted contact pipeline. All of it runs on free tiers.

The whole system

The pages are static MDX, built and served by Next.js on Vercel. The AI twin is a Node serverless function on the same deployment; it reads a search index that is built once at deploy time and bundled into the function, so there is no database to run. The game talks to a separate FastAPI service on a Hugging Face Space. Your browser only ever talks to this site and, for the game, that Space: Gemini is called server-side, and its key never ships to the client.

The parts

Each system has its own page now:

  • The AI twin: hybrid retrieval, a fallback chain of models, and honest failure modes.
  • The game backend: a PPO policy on CPU, one per difficulty, on a Space warmed before you notice.
  • The contact pipeline: a self-hosted form with a human check, one email out, and nothing stored.

Student Learning Assistant

A production RAG assistant that grounds each student in their own class recordings and materials. Handling student audio, so the deep write-up is shared on request.

AI twin

Hybrid retrieval, a fallback chain of models, and honest failure modes.