Pre-work · about 8 hours, released a month before

Get set up before day 1.

Four parts. Students arrive with a working setup, their Python and Git refreshed, and a first build with a coding agent.

Robert walks through the four parts and what to have ready on day 1. About two minutes.
Set up your machine1.5 hours

Paste it into Claude Code or Codex. It installs everything for you.

Done when: Your agent says uv, Python, Node, Git and GitHub are all set.

Python for AI apps2.5 hours
python-warmup % uv run pytesttest_refund.py .....5 passed in 0.35s
Where you'll end up

An agent wants to refund an order. Your coding agent writes the Python that checks it. Your job is to understand that Python: what a valid refund looks like, why "about fifty" gets stopped, and how the agent looks up an order.

curl -fsSL https://ai-engineering-course-outline.vercel.app/python-exercise.sh | sh
  1. Paste the command into Terminal. It puts the exercise in ai-course/python-warmup and runs the checks once, all red.
  2. Open that folder in your coding agent and ask: Fill in the three TODOs in refund.py, then run uv run pytest until it passes.
  3. Ask your agent to explain its three answers in plain words. Keep asking until you could explain them yourself: why a zero refund is rejected, why "about fifty" returns nothing, and why the tool waits for the order.

Done when: The checks say 5 passed, and you can say what each of the three parts does.

Git and GitHub1.5 hours
git-warmup % git push -u origin add-my-nameremote: Create a pull request for 'add-my-name' on GitHub by visiting:remote:   https://github.com/you/git-warmup/pull/new/add-my-name
Where you'll end up

You'll make the change every day of the course starts with: a branch, a commit, a push, and a pull request. One command sets up a practice repo on your GitHub; the Git commands are yours to type.

sh -c "$(curl -fsSL https://ai-engineering-course-outline.vercel.app/git-exercise.sh)"
  1. Paste the command into Terminal. It signs you in to GitHub, makes a private repo called git-warmup in ai-course, and pushes a first commit.
  2. In that folder, make a branch: git switch -c add-my-name
  3. Change hello.py so it prints your name, then save it as a commit: git add hello.py, then git commit -m "Print my name"
  4. Send the branch to GitHub: git push -u origin add-my-name
  5. Open the link Git prints and click Create pull request.

Done when: Your pull request shows as Open on GitHub.

Talk to an LLM like an engineer2.5 hours

Paste it into Claude Code or Codex in a new folder. Read every plan before you say go.

Done when: Your bill splitter's tests pass, it's committed, and you can explain what each file does.