Getting Started

Every module in the guide has three parts: an Overview that builds the idea and works the hand calculations the paper asks for, a Resources list of free readings, and a Checkpoint in the paper's own formats with hidden explanations.

Python & Data Preparation is a good example of the format.

You can do everything in Google Colab (Step 3), but a local setup is nicer for the many small pandas and scikit-learn exercises. We recommend uv for a fast, reproducible environment:

# Install uv (macOS/Linux)
curl -LsSf https://astral.sh/uv/install.sh | sh

# Create a project + install the essentials
uv init aio-practice && cd aio-practice
uv add numpy pandas scikit-learn matplotlib jupyter

Then uv run jupyter lab opens a local notebook server. Anything you rehearse here runs the same way in Colab on the day.

The applied paper is sat in Google Colab: one dataset, one notebook, a model file to export. Lessons that touch the pipeline carry a runnable Python box, so you can try the slice before opening a notebook.

Build the habit early: inspect, split, fit inside a Pipeline, then joblib.dump before you tune anything.

Every module has a status pill you can flip to Reading, Practicing, Complete, or Skipped. Progress is stored locally by default; sign in from the Dashboard to sync it across devices.

The Problems page is a filterable index of every question in the guide. Use it to find the blanks in one stage, or everything tagged leakage.

When a stage is done, sit a timed Round 1 mock. It is marked the way the paper is marked, and every question is explained after time is called. Current rules and dates are on the official site.