teaching

I am currently a part time assistant professor at École polytechnique, teaching CSC_43M02_EP Mining and learning on Web Graphs. In this class, students transform massive, real-world datasets into fascinating societal and scientific insights using advanced graph mining and Natural Language Processing. Recent projects have uncovered hidden structures across highly diverse domains. For instance, sin the realm of pop music, network analysis proved that a song’s sonic profile predicts its genre far better than an official genre predicts its sound, while mathematically identifying Stevie Wonder as the ultimate historical bridge between disparate musical lineages. By analyzing millions of European Parliament voting records, students discovered a cross-partisan “youth bloc” whose voting behavior transcends traditional political lines. Students also proved that companies frequently co-mentioned in financial news exhibit significantly higher stock return correlations, and showed that while scientific collaboration in AI happens in tight, localized clusters, intellectual influence and citations flow globally across these boundaries, with the highest impact uniquely rewarding multi-disciplinary research.