After obtention of my Ph.D in 2017 in computer science at IMT Atlantique (Brest, France), I joined EPFL (Lausanne, Suisse) as a postdoctoral researcher. There, I study (among others) interpretability of machine learning tools to extract meaningful features from complex data. In particular, I am involved in a project with the hospital of Lausanne (CHUV) in which we want to predict whether some treatments may affect positively patients suffering from a cancer, using features extracted from cell tissues.

My research interests mainly focus on graph signal processing (GSP), and more particularly on graph inference from signals, uncertainty principles and translations defined on the graph domain. These works, presented in my Ph.D. manuscript, allowed us to extend convolutional neural networks to have them applicable to signals evolving on irregular domain. In the continuity of these works, I am also interested in extending graph signal processing tools to multichannel signals (e.g., time series over a graph), using for instance EEG or fMRI signals.

During my Ph.D., I was very invested in teaching computer science. Jointly with other teachers and researchers, I have created a playful course, named PyRat, to initiate students to graph theory, algorithms and programming. This course was a resounding success with both the teachers and the students, and was adopted by the ECAM engineering school in Rennes, as well as by the school issued from the fusion of Télécom Bretagne and Mines de Nantes (currently IMT Atlantique).