JAX learning tracks

Follow a guided path or explore one topic. Pick up where you left off.

Start with the foundations

Learn JAX step by step

A guided sequence of lessons and exercises, from the basics onward.

Explore a topic

View problem sequences →

4 topic tracks

Build something complete

Turn JAX operations into a working experiment you can inspect and export.

Training data · two classes, shown as circles and squares

Mini-project

Build a Decision Boundary

1 exercise

A complete classifier project: train your model, inspect its learning curve, and export a runnable experiment.

Helpful to know

Tensor shapes, matrix multiplication, sigmoid and gradients.

Explore project
Sample images · horizontal and vertical patterns

Mini-project

Image Pattern Lab

1 exercise

Build a tiny image classifier using contrast filters and a learned classification head.

Helpful to know

Slicing, reductions, sigmoid and gradient descent.

Explore project
Causal mask · each token can attend only to itself and earlier tokens

Mini-project

Causal Attention Lab

1 exercise

Build a causal attention pipeline and inspect which earlier tokens each position uses.

Helpful to know

Matrix multiplication, stable softmax and attention shapes.

Explore project
Training pairs · noisy inputs (left), clean targets (right)

Mini-project

Denoising Autoencoder Lab

1 exercise

Train an encoder and decoder to reconstruct clean patterns from noisy inputs.

Helpful to know

Matrix multiplication, mean squared error and the chain rule.

Explore project