Course overviewYour accountSign in

Machine learning from first principles, with a tutor that never runs out of patience.

A graduate-level, hands-on course — 14 chapters from the ML workflow to transformers, with open-ended quizzes you can't bluff.

Start with a free lesson →
A wireframe loss landscape with a basketball resting high on one slope: the loss L of one network at θ̂ + α₁d₁ + α₂d₂, its trained parameters moved along two random directions in weight space; axes labelled α₁, α₂ and loss.

Height is the loss: how wrong one network is as its weights change. Training rolls the ball downhill to the best weights it can find. That is gradient descent.

What his students said

wrote this course. He is a Continuing Instructor at UC Berkeley, where he teaches applied machine learning. Uriversity is new, so nobody has finished it yet. These are anonymous course evaluations from Uri's students at UC Berkeley.

He can make students grow their own thinking power.
Course evaluation · Spring 2026
Making everyone feel included, a safe learning space, and always being open for questions was a huge strength, and personally very appreciated.
Course evaluation · Spring 2026
It's clear that he is constantly exploring new and creative ways to present the material more effectively.
Course evaluation · Summer 2026

Join the next cohort

Chapters 1 and 2 are free for anyone to read.

Applied Machine Learning
New here? The course overview → — what you'll learn, what you need, and what a week asks of you.
The full 14-chapter arc. Chapters open on a weekly schedule — see the schedule → · full syllabus → · glossary → · references →
Loading lessons…