Skrub - Machine Learning with data frames

​Speaker : Riccardo Cappuzzo, research engineer and lead developer of skrub

​How can you turn complex, heterogeneous tables into features ready for machine learning?

​In partnership with P16, an Inria-led initiative, we invite you to an online presentation of skrub, a Python library that simplifies preprocessing and feature engineering for tabular machine learning.

​Compatible with scikit-learn, pandas and Polars, skrub helps you prepare datasets that mix numerical values, categories, dates and text, and combine information from multiple tables. It makes it easier to build reusable pipelines—from data preparation to prediction—with less manual code.

​Whether you are working with database tables, encoding categorical and text features, or building a first classification or regression model, this webinar will introduce skrub’s key capabilities and how to integrate them into your machine learning pipelines.