Local Development Setup

Some challenges require dependencies that can't run in the browser. The CLI tool sets up a local workspace and runs the test suite for each challenge.

1. Install the CLI

pip install ai-deep-dive

This installs the ai-deep-dive command globally. Requires Python 3.8 or later.

2. Initialize the course

ai-deep-dive init build-chatgpt

Creates a workspace folder with starter files for every challenge, organized by chapter.

build-chatgpt/
01-from-text-to-bytes/
01-01_utf8_encode.py
01-02_utf8_decode.py
02-tokenization/
...

Use --dir my-folder if you want a different folder name.

3. Work on a challenge

cd build-chatgpt
nvim .  # or your preferred editor

Open the folder in your editor and find the challenge file. Each filename starts with the challenge ID, for example 01-02 means Chapter 1, Challenge 2.

4. Test your solution

ai-deep-dive test 01-02

Runs the test suite for that challenge. The CLI expects the filename to follow the naming convention and searches the current directory recursively, so you can organize your folders however you prefer.

5. Sync your progress

ai-deep-dive sync

Automatically imports locally completed challenges to the website. Progress is stored in your browser on this machine.

Other commands

statusShows which challenges you've completed locally
listLists all available courses

FAQs

Back to Course