My name is Maxime Lagrange!

I am a doctoral researcher in physics at the Centre for Cosmology, Particle Physics and Phenomenology (CP3) of UCLouvain. My work lies at the intersection of experimental particle physics, computational modeling, and artificial intelligence. I specialise in the simulation, data analysis, and optimisation of muon tomography systems — an emerging imaging technique that uses cosmic rays to probe dense or large-scale objects.

Over the past few years, I have developed and maintained several research software libraries in this field, including Muograph, a Python-based framework for muon scattering data analysis, and TomOpt, a PyTorch-powered library for the differentiable optimisation of detector systems. My research integrates physical simulation (GEANT4 and fast Python-based models), deep learning architectures (DNNs, 3D CNNs, Transformers, RNNs), and differentiable programming techniques to build fully end-to-end systems for experimental design and data inference.

My broader interest is in the convergence of physics and machine learning — using AI not just to analyse data, but to design the instruments that collect it.

• Muon Scattering & Absorption Tomography — detector design, simulation, and reconstruction
• Differentiable Programming — end-to-end optimization of physics instruments
• Deep Learning for Physical Inference — momentum estimation, image denoising, and data-driven modeling
• Monte Carlo Simulation — GEANT4, Python-based fast scattering models
• Scientific Software Engineering — open-source tools for physics data analysis and detector optimization

• Toward Using Cosmic Rays to Image Cultural Heritage Objects
  iScience, 2025 — Read Here
• TomOpt: Differential Optimization for Detector Design in Muon Tomography
  Machine Learning: Science and Technology, 2024 — Read Here
• Toward the End-to-End Optimization of Particle Physics Instruments
  Reviews in Physics, 2025 — Read Here