AI Roadmaps
← All roadmaps
BeginnerRoadmap 01 · 13 steps · 6h 54m of video

Python & Math for AI

The Python, NumPy, linear algebra, calculus, and probability you actually need before touching models.

Before this: None. Start here if you have never coded.

  1. 2h 02m
    Step 1

    Python from zero

    Variables, loops, functions, and classes in one sitting.

    Python Full Course for Beginners

    Programming with Mosh

  2. 58 min
    Step 2

    NumPy arrays

    Every ML library speaks NumPy. Learn arrays, slicing, and broadcasting.

    Python NumPy Tutorial for Beginners

    freeCodeCamp.org

    rougier/numpy-100★ 14.7k
  3. 1h 34m
    Step 3

    Pandas for data

    Load, filter, group, and clean real tables.

    Complete Python Pandas Data Science Tutorial! (2025 Updated Edition)

    Keith Galli

    guipsamora/pandas_exercises★ 13.4k
  4. 31 min
    Step 4

    Plotting with Matplotlib

    If you can't plot your data, you can't debug your model.

    Learn Matplotlib in 30 Minutes - Python Matplotlib Tutorial

    Tech With Tim

    jakevdp/PythonDataScienceHandbook★ 50.5k
  5. 10 min
    Step 5

    Vectors

    The building block of embeddings, weights, and everything else.

    Vectors | Chapter 1, Essence of linear algebra

    3Blue1Brown

  6. 11 min
    Step 6

    Matrices as transformations

    The single best intuition for what a neural-network layer does.

    Linear transformations and matrices | Chapter 3, Essence of linear algebra

    3Blue1Brown

  7. 10 min
    Step 7

    Matrix multiplication

    Why stacking layers is composing functions.

    Matrix multiplication as composition | Chapter 4, Essence of linear algebra

    3Blue1Brown

  8. 17 min
    Step 8

    Eigenvectors and eigenvalues

    The idea behind PCA and many stability tricks.

    Eigenvectors and eigenvalues | Chapter 14, Essence of linear algebra

    3Blue1Brown

  9. 17 min
    Step 9

    The essence of calculus

    Derivatives and integrals, visually, without the textbook pain.

    The essence of calculus

    3Blue1Brown

  10. 17 min
    Step 10

    Derivatives

    Gradients are just derivatives, and gradients train every model.

    The paradox of the derivative | Chapter 2, Essence of calculus

    3Blue1Brown

  11. 15 min
    Step 11

    Bayes' theorem

    How to update beliefs with evidence. Shows up everywhere in ML.

    Bayes theorem, the geometry of changing beliefs

    3Blue1Brown

  12. 5 min
    Step 12

    The normal distribution

    The distribution behind initialization, noise, and diffusion.

    The Normal Distribution, Clearly Explained!!!

    StatQuest with Josh Starmer

  13. 6 min
    Step 13

    Maximum likelihood

    The principle behind almost every loss function.

    Maximum Likelihood, clearly explained!!!

    StatQuest with Josh Starmer

    mml-book/mml-book.github.io★ 16.2k

Build it

Watching isn't learning. Pick one and put it on GitHub.

Where to go next