Skip to content

Getting Started

This page is only for local setup and the first working run.

If you need route guidance, use Beginner Path or IOAI Path. If you want the teaching philosophy, use Academy Model.

Before You Install Anything

You should already be able to:

  • open a terminal
  • run a basic Python command
  • create a virtual environment
  • read a short error message without panicking

If those steps are still hard, fix that first. AI Academy is beginner-friendly, but it is not a zero-computer-literacy course.

Quick Setup

From the noki-website/ directory, run:

python3 -m venv academy/.venv
academy/.venv/bin/python -m pip install -r academy/requirements-docs.txt
academy/.venv/bin/python -m pip install -r academy/requirements.txt
academy/.venv/bin/python -m mkdocs serve -f academy/mkdocs.yml --dev-addr 127.0.0.1:8001

Then open http://127.0.0.1:8001/.

Use 8000 instead if it is free on your machine. 8001 is the safer fallback when another local service is already listening on 8000.

First Working Run

After the docs load, confirm the environment with one small example:

academy/.venv/bin/python academy/examples/course-support-inspection/quick_inspection.py

If that works, your first real academy loop is available.

First 60 Minutes

If you are new to AI:

  1. read Beginner Path
  2. read Table Inspection
  3. run academy/.venv/bin/python academy/examples/numpy-shape-sanity/shape_sanity.py
  4. run academy/.venv/bin/python academy/examples/course-support-inspection/quick_inspection.py
  5. write one short note about the first failure pattern you noticed

If you are preparing for IOAI:

  1. read IOAI Path
  2. read Honest Splits and Baselines
  3. run academy/.venv/bin/python academy/examples/validation-baseline-comparison/baseline_comparison.py
  4. open Decision Clinics
  5. write one stop-or-continue note before you move into a full track

What To Inspect On The First Run

Do not just check that the command finished. Inspect:

  • the table shape
  • the column names
  • missing values
  • target balance or label counts
  • anything that looks suspiciously informative

The academy is not using code correctly if the learner only notices that a script "ran."

Common First-Run Problems

If imports fail:

  • use academy/.venv/bin/python, not the system python3
  • reinstall with academy/.venv/bin/python -m pip install -r academy/requirements.txt

If MkDocs does not start:

  • make sure requirements-docs.txt is installed
  • switch from port 8000 to 8001

If a script cannot find a file:

  • run it from noki-website/
  • keep the current working directory there when using the academy/... paths above

What To Do Next

After the first working run:

  1. pick your route: Beginner Path or IOAI Path
  2. use Examples for short runs
  3. move to Tracks when you want a full workflow