AI Engineering with
Professor Synapse

Retrieval practice is how lessons stick — not re-reading, not passive watching. Professor Synapse fires the concept at you and makes you apply it on the spot. Before he explains anything, he asks what you already think. That's not a personality quirk. That's the whole point.

This runs in Claude Code. Every session is scoped to your current phase and progress — nothing else. No inflated system prompts. The design is intentional: do more per token, not more tokens.

Synapse doesn't teach from one persona. When you're in the transformers phase, he loads a transformer specialist. Deployment phase — different expert. He's always relevant because he's never generic.

Before lesson one, Synapse runs a placement check. You enter where your knowledge actually is, not where someone assumed you'd start. No wasted sessions grinding through things you already know.

Codecademy with a chatbot doesn't build understanding, it builds dependency. A tool that explains the answer doesn't teach you to find it.

Build a calculator. Build a to-do app. Build something you'll never use. Every capstone here connects to what you're actually working on.

AI moves quarterly. Most courses were written for a market that no longer exists. This curriculum is reviewed against what's shipping, not what shipped.

A dozen sources, zero coherence. Browser graveyards and no mental model. One path. One sequence. One place to be.