AI FDE Handbook

Appearance

00.4 — Prerequisites ​

This handbook does not teach Python syntax, basic programming concepts, or "what is an API" from zero. It assumes:

Required going in ​

  • Python fluency: comfortable with functions, classes, decorators, context managers, exceptions, virtual environments, and reading someone else's codebase. You don't need to have memorized asyncio internals — Part 1 covers that in depth — but you should not be learning what a list comprehension is from this book.
  • Basic command line comfort: navigating directories, running scripts, reading a stack trace, using git for at least add/commit/push/branch/merge.
  • Has used an LLM API directly (OpenAI, Anthropic, or similar) — sent a prompt, gotten a completion back, ideally used streaming or function calling at least once.
  • Has used LangChain, LangGraph, and/or LangSmith at a basic level — built at least one chain, one graph, or looked at one trace. This book explains their architecture and internals, not their basic syntax; if you've never opened LangGraph before, expect Part 5 to take noticeably longer than someone who has.

Helpful but not required ​

  • Prior exposure to REST API design, SQL, and relational databases (Part 1 covers these, but from a "production patterns" angle rather than "what is a JOIN").
  • Some exposure to Docker (Part 7 assumes you can read a Dockerfile even if you've never written a production one).
  • Any experience in a customer-facing or cross-functional role (support, sales engineering, consulting) — this maps directly onto FDE instincts, even if the domain was different.

Not required ​

  • A deep ML/statistics background. Part 2 covers ML/DL fundamentals at the depth an AI engineer actually needs (understanding what's happening, not deriving backpropagation from scratch).
  • Prior enterprise software experience. Part 10 and Part 12 build this from first principles.
  • Any specific cloud provider certification. Part 7 covers cloud concepts generically with concrete examples, not a certification syllabus.

Environment you'll want set up ​

  • Python 3.11+ with uv for dependency management (this book's code examples assume uv).
  • A code editor with a Python language server.
  • API keys for at least one LLM provider (Anthropic or OpenAI) — free-tier/low-spend is enough for every exercise in this book.
  • Docker installed, for Part 7 onward.
  • A GitHub account, for Part 1's Git section and for the project work in Part 16.

If any of the "Required going in" items are shaky, it is worth spending a week closing that gap before starting Part 1 — not because this book won't explain those concepts when they come up, but because it will move quickly past them to get to the material that's actually differentiating.