About
I am an Assistant Professor (Maître de Conférences) at the Laboratoire de Mathématiques d'Orsay of Université Paris‑Saclay, and a member of the Inria Celeste team.
Previously, I was a postdoctoral researcher at EPFL (BAN chair) in 2022-2023. I completed my PhD in 2022 at Inria Paris, supervised by Laurent Massoulié and Marc Lelarge.
Research interests
- Fairness and causality
- Optimal transport for statistical learning
- Statistical inference in graphs and matrices
- Information/computation thresholds on random instances
- Opinion dynamics and misinformation in networks
News
📝 December 2025: a new preprint is out! Together with Bertrand Even, we study the alignment of multiple correlated Gaussian graphs and reveal a statistical–computational gap. We first establish the informational threshold of the problem when the number of graphs p grows with n, the number of nodes. Interestingly, while a larger p initially helps the informational threshold, beyond a certain point it no longer eases the problem: the difficulty reduces to aligning a single graph with the parent graph. In the very large p regime, the partial and exact recovery thresholds no longer coincide (the all-or-nothing phase transition disappears). Finally, we show a computational barrier in the low-degree framework: when the correlation ρ is smaller than 1/polylog(n,p), low-degree polynomials can no longer extract signal, giving a clear picture of the stat–comp gap in this problem. See the preprint.
🎓 I am looking for a PhD student. Together with Evgenii Chzhen, we have a PhD offer at the interface between mathematical statistics and machine learning, namely about Handling unfairness in data: modelling, detecting, and debiasing. Have a look at the PhD proposal. Students are encouraged to apply by email.
📝 September 2025: a new preprint is out! Together with Lucas De Lara, we investigate the connections between statistical transport maps and causal inference. Our work compares three notions of multivariate monotone transport—cyclically monotone, quantile-preserving, and triangular monotone maps—deriving conditions for their equivalence and clarifying their structural properties. We then show how counterfactual reasoning in structural causal models can be formulated as a problem of selecting transport maps, highlighting when causal assumptions align with classical statistical transports. See the preprint.
Job offers
🎓 I am looking for a PhD student. Together with Evgenii Chzhen, we have a PhD offer at the interface between mathematical statistics and machine learning, namely about Handling unfairness in data: modelling, detecting, and debiasing. Have a look at the PhD proposal. Students are encouraged to apply by email.