Explore tabular learning and foundation models in this 22-minute talk by Gaël Varoquaux, research director at Inria and co-founder of scikit-learn & Probabl. Discover how tabular data, despite being central to organizations, has been overlooked in AI discussions focused on images, text, and sound. Learn about the challenges of data wrangling in tabular data and how Varoquaux's approach aims to simplify this process. Gain insights into skrub, a new library that facilitates complex tabular learning pipelines and automates wrangling operations. Understand how a few lines of skrub code can replace dozens of manual wrangling lines, potentially revolutionizing your data science workflow.