Alix Leroy

I am an applied mathematician, a passionate scientist and a feminist.

I am currently doing my PhD in Applied and Computational Mathematics on efficient Bayesian approaches to inverse problem. I am supervised by Jonas Latz, Des Higham and Ben Leimkuhler.

I work on:

• Adaptive stepsize Langevin dynamics to obtain sampling methods that are more computationally efficient, see my first paper.
• Graph-based learning methods for semi-supervised classification, where Langevin dynamics are used to infer labels together with uncertainty quantification. The approach performs well even with very few labeled nodes. (Manuscript in preparation, June 2026.)
• Ensemble Kalman optimisation methods, building on Ensemble Kalman Inversion with a novel modification. The resulting algorithms are inexpensive, derivative-free, and adjoint-free, and are suitable for large linear inverse problems. We establish convergence guarantees. (Manuscript in preparation, June 2026.)

I receive the support of the Maxwell Institute Graduate School in Mathematical Modelling, Analysis and Computation CDT (MAC-MIGS) and the Luxembourg National Research Fund (FNR).

I have two BSc from the Universite Libre de Bruxelles (ULB), in physics engineering in 2017 and business management in 2015. I also received an MSc in Environment and development in 2018 from the London School of Economics(LSE). I worked as a researcher in DRC (Democratic Republic of the Congo) in development economics and in London for the What Works for Children Social Care, managing and evaluating RCTs (Randomised Controlled Trials). I finally enrolled in an MSc in Applied Computational Mathematics at the UoE which led me to my current PhD.

Life is a journey, you might as well enjoy it, learn as much as you can and have a positive impact along the way!

Representation matters, please get in touch if you would like to get a neurospicy and feminist take on doing a PhD in applied math!