I am currently working as a senior data scientist at the Swiss Data Science Center in ETHZ. My current focus in machine learning is deep neural newtorks and generative models (autoencoders and GANs).
I attach a strong importance to reproducible research.
I did my PHD in the LTS2 laboratory of EPFL under the supervision of Pierre Vandergheynst. I specialized in signal processing and (convex) optimization but because of my interest in electrical networks, I did a minor in energy. The purpose of my thesis is to establish a strong basic theory for graph (network) signal processing. In particular, I studied graph uncertainty principles, graph-manifold consistency results and graph learning algorithms. Although some of my work is very theoretical, I tried to always keep in mind the possible practical applications. I was especially involved in developing different toolboxes for spectral graph theory algorithms (GSPBox) and for convex optimization (UNLocBoX).
I was born in 1987 in Le Cotterg, a quiet village in the heart of the Swiss Alps. After graduating from the College of La Royale Abbaye de St-Maurice, I’ve decided to study electrical engineering at EPFL, the Swiss Federal Institute of Technologies (Lausanne). Thanks to my Bachelor degree results, I was granted an “Excellence Fellowships”, an award reserved to the best student of each EPFL section. In 2012, I finished my Master with a specialization in signal processing and a minor in energy. I then studied for six months at the Acoustic Research Institute of Vienna (ARI) where I worked on Gabor frames and phase reconstruction. In 2013, I started my thesis in the LTS2 laboratory of EPFL under the supervision of Pierre Vandergheynst. My Ph.D. focused on spectral graph theory, machine learning, optimization and audio signal processing. In 2015, I organized a summer school in Leukerbad named "Key Insights on Networks and Graphs". After graduating, I started to work as a senior data scientist in the Swiss Data Science Center, where I specialized myself in various areas of deep learning such as generative adversarial networks.