AI Engineering — Start Here
A knowledge base for becoming an AI applications engineer, built around one anchor project: the Data Patch Agent. It has two halves that stay wired together:
- 📚 Knowledge Bank — what I should know, as generic theory: foundations → intermediate → expert.
- 🧪 Learnings — what I actually did: the decisions I made, things I discovered, mistakes I fixed, and reusable patterns.
The two link both ways: each learning links up to the concept it relates to, and a concept links down to my work via a Related implementations line.
The loop
- Build something (on the Data Patch Agent or elsewhere) and record it in Learnings.
- Tell the assistant what you learned, decided, or fixed.
- It checks the reference and either adds new theory (researched, with sources) if the topic is missing, links your implementation if it aligns, or flags it if your approach diverges from better practice — per ASSISTANT_PROMPT.md.
Read in this order
- Knowledge Bank · Overview — the map, and how the two halves connect.
- 01 · Foundations — the beginner concepts.
- 02 · Intermediate and 03 · Expert — build and production concepts, as you reach for them.
- Keep the Glossary open for quick term lookups.
Companion documents
| Document | Type | Use it when… |
|---|---|---|
| Knowledge Bank | The theory | You want the structured "what I should know" |
| Learnings | The practice | You want the real record of what you built and why |
| Glossary | Reference | A term trips you up |
| Building an Agentic System | How-to | You're building an agentic system from scratch |
| AI-First Methodology | Reference | Understanding the six components of an AI-first practice |
| Reading List | Sources | You want the essays behind the book |
| Open Questions | Inbox | Something isn't clear yet and needs revisiting |