Louis Berthier

I am a PhD student at the Centre de Mathématiques Appliquées (CMAP) of École Polytechnique, in the SIMPAS team. My research is a CIFRE collaboration with Michelin, supervised by Pr. Eric Moulines, Dr. Ahmed Shokry, Dr. Maxime Moreaud, and Guillaume Ramelet.

I build machine learning systems that predict tire quality in real time on Michelin’s production lines, replacing slow and costly laboratory measurements, and, just as important, that flag when their own prediction should not be trusted. My work sits at the crossroads of machine learning, uncertainty quantification (especially conformal prediction), and explainability, with the long-term goal of enabling autonomous process control in manufacturing. You can read more on my research page.

Before my PhD, I explored machine learning across several domains:

• Robotics: Adaptive & Intelligent Robotics Lab, Imperial College London, sample-efficient Quality-Diversity optimization with surrogate models
• Neuroscience: CNRS CerCo, deep learning for pathological oscillation detection in epileptic EEG signals
• Healthcare: EuroMov DHM, image processing and biomedical signal analysis

Collaboration

I enjoy bridging academic research and real-world applications. Alongside my PhD, I work with startups and industry partners on applied ML, from agentic systems to computer vision at scale to uncertainty-aware prediction.

If you are working on a problem where machine learning has to be reliable, not just accurate, I would genuinely like to hear about it. That can take many shapes, a research internship, focused consulting, or a joint project with concrete deliverables.

Domains of interest:

• AI for Science: Neuroscience, Bio AI, Healthcare
• Industry: Manufacturing, Robotics, Supply Chain, Spatial & Open-world AI
• Commerce: Retail, Finance

Interested? Let’s talk.