About Me
I am a postdoctoral researcher in computer science at the Laboratoire Méthodes Formelles (LMF, UMR 9021). I am currently working on the inference of chemical reaction networks (ODE systems) from time series observations from bioreactors. This work combines deep neural networks with formal constraints to ensure consistency, interpretability, and robustness.
I’ve done my PhD in the Dyliss team at IRISA (UMR 6074) on the inference of Boolean regulatory rules controlling bacteria metabolism. To address this problem, I’ve developed an optimization modulo theory frameworks combining logic and quantified real linear arithmetics constraints.
Research Interest
Keywords: constraint programming (OMT) – machine learning – dynamic systems – formal methods – bioinformatics
My research lies at the intersection of formal methods, hybrid optimization, and machine learning. It is motivated by inverse problems in bioinformatics that are inherently computationally challenging due to the heterogeneity and noise of experimental data, as well as the hybrid nature of biological systems dynamics (continuous and discrete at different timescales). I’m particularly interested in developing novel hybrid inference frameworks leveraging the formal guarantees of constraint programming with the predictive power and scalability of machine learning.