Math for AI
The math behind the magic — no PhD required
Linear algebra, calculus, and probability taught as intuition — exactly the math you need to understand machine learning, nothing you don't.
Beginner3 modules · 12 lessons~91 min855 XPAligned with 3Blue1Brown · MIT OpenCourseWare
Progress0/12
Module
Linear Algebra
The geometry every model secretly runs on
Vectors: Arrows and Lists7 min · 60 XPVectors: Arrows and Lists
Matrices Are Verbs7 min · 65 XPMatrices Are Verbs
Dot Products Measure Agreement8 min · 70 XPDot Products Measure Agreement
Matrix Multiplication Is Composition8 min · 75 XPMatrix Multiplication Is Composition
Eigenvectors: The Directions That Survive8 min · 80 XPEigenvectors: The Directions That Survive
Module
Calculus for ML
Derivatives, gradients, and the chain rule — the machinery that lets models learn
Derivatives: The Slope at a Point7 min · 65 XPDerivatives: The Slope at a Point
Gradients: The Compass of Steepest Ascent8 min · 75 XPGradients: The Compass of Steepest Ascent
The Chain Rule: Why Backprop Exists8 min · 80 XPThe Chain Rule: Why Backprop Exists
Partial Derivatives & Jacobians6 min · 60 XPPartial Derivatives & Jacobians
Module
Probability for ML
Uncertainty as a tool — distributions, expectation, and updating on evidence
Probability Is a Belief with a Number7 min · 65 XPProbability Is a Belief with a Number
Distributions, Expectation & Variance8 min · 75 XPDistributions, Expectation & Variance
Bayes' Rule: Belief Updating9 min · 85 XPBayes' Rule: Belief Updating
Capstone project
The geometry of learning
Implement a tiny gradient-descent stepper from scratch in Python — no libraries for the math. Visualize how different learning rates behave on a simple function.
Deliverable: A notebook with your from-scratch implementation plus plots comparing learning rates.
Milestones
Finish every lesson to earn the Mathematics for AI Certificate.