👋 Welcome !

I am Perceval Beja-Battais, currently a PhD student in applied mathematics at Centre Borelli, supervised by Nicolas Vayatis.
I work on the interface between Machine Learning, optimization and dynamical systems with an industrial applications to Model Predictive Control in nuclear engineering. My thesis is carried out within the framework of a CIFRE collaboration between academia and industry, in partnership with Framatome, and focuses on the development of new approaches for the simulation and control of complex physical systems.

My research interests include physics-informed neural networks (PINNs), optimal control, uncertainty quantification, and the analysis of stability for nonlinear dynamical systems. I am particularly interested in bridging classical methods from numerical analysis (e.g. collocation, adjoint-based optimization) with recent advances in Machine Learning to design hybrid algorithms that are both interpretable and scalable.

Prior to my PhD, I graduated from Mines Nancy, ENS Paris-Saclay (MVA) & Sorbonne Université, where I specialized in applied mathematics (See CV).

In the longer term, I am interested in exploring how advanced learning-based methods can be integrated into industrial decision-making processes, both in energy and beyond, and I am open to interdisciplinary collaborations at the interface between mathematics, physics and computer science.