Volume 1 · Eight-week foundations

Mathematical Foundations for Machine Learning and AI

Algebra, mathematical notation, linear algebra, calculus, and NumPy — rebuilt carefully for experienced engineers who want to read papers, follow derivations, and implement models from scratch.

20
chapters
10
interactive labs
320+
exercises & solutions

Who this is for

You already program professionally and understand algorithms and software architecture. What you want is mathematical fluency: the ability to open a paper and read every symbol, follow the derivation, and reproduce the result in code. This course treats you as an engineer, not a beginner — careful explanations, no hand-waving, and every concept connected to implementation.

What you'll be able to do

  • Read ML and AI papers without stalling on notation
  • Follow derivations and reproduce important equations
  • Translate mathematics directly into correct NumPy
  • Implement algorithms and reproduce experiments from scratch
  • Spot hidden assumptions and dimension errors
  • Understand why training works — gradients, losses, optimization

How the mathematics connects to ML

Linear algebra → architecture

A neural layer is Wx + b. Shapes, matrix products, and rank are the language of every model — from embeddings to attention.

Calculus → learning

Training is optimization. Derivatives measure sensitivity; the chain rule is backpropagation; gradients point the way downhill.

NumPy → implementation

Every equation becomes vectorized code. You'll translate summations, dot products, and gradients into correct, shape-checked NumPy.

The eight-week plan

Full curriculum →
  1. Week 10%

    Algebra and Mathematical Language

    Reading and unpacking the symbols ML papers assume you know.

  2. Week 20%

    Functions and NumPy Foundations

    Models and losses are functions; NumPy is how we evaluate them.

  3. Week 30%

    Vectors

    The geometry and algebra of feature vectors and embeddings.

  4. Week 40%

    Matrices and Linear Systems

    Matrices as data, as operations, and as transformations of space.

  5. Week 50%

    Vector Spaces and PCA Intuition

    Directions, dimensions, variance, and dimensionality reduction.

  6. Week 60%

    Limits and Derivatives

    Rate of change as the foundation of all learning.

  7. Week 70%

    Chain Rule and Gradients

    How gradients are computed through composed functions.

  8. Week 80%

    Optimization and Linear Regression

    Assembling calculus, linear algebra, and NumPy into a learner.

The five parts

  1. Part 1
    Algebra and Mathematical Notation
  2. Part 2
    Functions and Graphs
  3. Part 3
    Core Linear Algebra
  4. Part 4
    Calculus
  5. Part 5
    NumPy Laboratory

How to use this course

Lessons

Each chapter builds from motivation and intuition to a formal definition, a worked example, an ML use case, runnable NumPy, and common mistakes.

Laboratories

Interactive visualizations for vectors, matrices, derivatives, gradient descent, and PCA — manipulate the math and watch it respond.

Exercises

Four difficulty levels from recall to research-thinking, with graduated hints and complete, gated solutions you reveal deliberately.

Everything ends in a capstone: reproducing linear regression from scratch, from the equations to a gradient-checked implementation and a short research report.