Applying Elixir to Data Science: the Nx Ecosystem from Zero to Hero (virtual)

Paulo Valente



This tutorial is a zero-to-hero course on the Nx ecosystem, a numerical computing family of libraries used to integrate typed, multidimensional data needed for machine learning and data science. We will be creating a solution from scratch for one of the Kaggle “Getting Started” competitions, building all of our code using Nx tooling. The Kaggle “Getting Started” competitions are problems and challenges, with associated data sets, aimed at growing your data science skills.Throughout the training, we will make the boundaries between the Nx libraries clearer, allowing us to discuss how they complement and interact with each other, ensuring you become an expert.

Training objectives:

  1. Understanding the Nx Ecosystem of libraries
  2. Understand how the libraries complement and interact with each other
  3. Get an introduction to Machine Learning in Elixir using Nx

Training audience: This training will be interesting both to people who have some experience in ML, little experience in Elixir, but want to learn Nx, as well as those who know Elixir, have little ML knowledge and want to see how the Nx libraries fit together.

Training prerequisites: Participants should make sure they can run EXLA on their machines. Ideally, CUDA should be enabled, but CPU should suffice for the simpler models. The latest Livebook should also be installed.

Experience level:
Intermediate Users