AI Engineering from Scratch: choose a route through the curriculum
AI Engineering from Scratch: choose a route through the curriculum
Use one learning goal and a runnable artifact to find a sensible entry point in this large course.
What you will learn
- Start with the work you want to do
- Treat the lesson artifact as evidence
- Know the boundary of this review
Before you start
- Basic command-line access
- A bounded learning goal and disposable checkout
Turn lesson commands and code changes into a reviewable record of what you can do.
Key takeaways
- The repository is a course rather than a deployable AI service.
- Entry points depend on the learner’s goal and prerequisites.
- A saved execution record makes learning claims inspectable.
Start with the work you want to do
The repository is a curriculum, not an agent runtime. Its README maps beginners to setup, model builders to math and machine learning, and application developers to LLM or agent engineering. The advertised lesson count is a description of the upstream repository, not a measure of learning quality.
A reader who already writes Python need not begin with terminal basics. Pick a route, then inspect its prerequisites and the first lesson before committing to a long sequence. The placement skill is optional; the lesson files remain readable without an AI tutor.
Treat the lesson artifact as evidence
The project asks readers to read a lesson, type its important code, run it from the repository root and keep the command, exit code and output. That is more useful than merely marking a page complete.
For a first exercise, run the environment preflight and one dependency-free linear algebra example. Save what the script printed and explain one change you made. If the result surprises you, follow the code before asking an agent to summarize it.
Know the boundary of this review
The course covers models, protocols, agents, infrastructure and capstones across multiple phases. This series examines a fixed source revision and a few representative lessons; it does not certify every exercise or claim to have completed the curriculum.
Later chapters show how to set up a local study environment, read the curriculum structure and evaluate your own progress without treating the website or the optional tutor as a substitute for runnable work.
Decision guide
| Criterion | Option A | Option B |
|---|---|---|
| Best when | You need predictable behavior and easy auditing | You need adaptive optimization and have reliable telemetry |
| Main risk | May leave performance on the table | Can become difficult to explain or debug |
Implementation steps
- 1
Choose one goal from the README route table.
- 2
Run the environment preflight in a disposable checkout.
- 3
Save one command, its exit code and a changed artifact.
Copy-ready example
python3 phases/00-setup-and-tooling/01-dev-environment/code/verify.py --route beginner
python3 phases/01-math-foundations/01-linear-algebra-intuition/code/vectors.pyFrequently asked questions
Must I read every lesson in order?
No. The README provides goal-based routes, though later lessons may assume earlier concepts.
Is the AI tutor required?
No. The lessons and code can be read and run directly from the repository.
Sources
- AI Engineering from Scratch / README.mdSource checked 2026-09-29
- AI Engineering from Scratch / phases/00-setup-and-tooling/01-dev-environment/docs/en.mdSource checked 2026-09-29