Research interests

My research revolves around numerical linear algebra and high performance computing.

Since 2014, I investigate the use of low-rank approximations within direct solvers to reduce their computational cost. I am particularly interested in the complexity analysis of such methods, as well as their high performance implementation on parallel computers.

More recently, as postdoc @ Univ. of Manchester, I have started investigating floating-point arithmetic and in particular the error analysis of numerical linear algebra algorithms.

Please refer to my publications or talks pages for a list of my recent research activities.


Current projects

Past projects

PhD thesis

Title: Block Low-Rank multifrontal solvers: complexity, performance, and scalability.
Advisors: Patrick Amestoy and Alfredo Buttari.
Date of award: 24 November 2017.
Thesis manuscript is available here.
I received the 2018 Gilles Kahn prize for my thesis.
See also this article I wrote for the popular science blog Binaire hosted by the French newspaper Le Monde.