In this tutorial I introduce JAX, the new google framework for deep learning. We first introduce the main peculiarities and describe how to compute gradients to optimize simple ML algorithms. Eventually, we implement more complex NLP pipelines, showing were JAX modularity comes to shine.