David Saulpic


About me

I am a researcher at IRIF ("Chargé de recherche"), affiliated with CNRS and Université Paris Cité.
I was previously a Marie-Sklodowska-Curie fellow at IST Austria, hosted by Monika Henzinger. I did my PhD in Sorbonne Université, in Paris, where I was very glad to be have been advised by Vincent Cohen-Addad and Christoph Dürr. My thesis received the Gilles Kahn award.

I am broadly interested in algorithms, with a particular attention to Clustering problems for which I strive for a clear picture of their multiple facets: for instance, when is it possible to recover clusters, with what precision, how much data is necessary?
I am also trying to understand the success of k-Means objective function among practitioners. Is it because we love to assume we live in a Gaussian world? Or because of Lloyd's algorithm? If you have an idea on that matter, I'd love to discuss it with you!

Besides clustering, I am more generally interested in theory of problems arising from data analysis. I would like to design algorithms for those problems, with provable guarantee, for instances with constraints on the memory usage, privacy or robustness requirements.

I am actively doing popular science, trying to share knowledge from TCS to wide audiance. Duing my PhD, I worked at la Cité des sciences et de l'industrie, in Paris, and am now designing an animation on Privacy. I will present it soon at the festival Pint of Science!

Contact


You can find a (hopefully up-to-date) CV here

Publications