AI Engineering Foundations

Who this series is for

This series is for developers moving into AI engineering from another background: data analysts who write some Python, frontend engineers who have never deployed a backend service, or recent graduates who know algorithms but not production tooling. Most failed AI projects do not fail at the model. They fail at the plumbing around it: an API that times out, a container that behaves differently in production, a database query that falls over at ten thousand rows. That plumbing is what this series covers, deliberately before any LLM appears.

What each chapter covers

• Python & APIs covers how to call and build HTTP APIs with FastAPI, how JSON serialisation and validation work, and why typed request and response models save you from a whole class of bugs that show up later when an LLM starts producing your inputs.
• Data & Databases covers relational modelling, SQL you will actually use (joins, indexes, transactions), and how to reason about when a query is slow. The same PostgreSQL skills carry straight into pgvector in the Production RAG series.
• Docker & Deployment covers building reproducible images, environment configuration, and the difference between something that works on your laptop and something that runs reliably on a server.

How to approach it

If you already ship backend services professionally, skim the chapter checklists and move on to Building LLM Applications. If any of the checklist items are unfamiliar, work through the chapter and do the exercises by hand rather than copying the code. Each chapter is designed to take one to two hours, and the whole series is roughly six hours of focused work.