Quentin Guilhot

Quentin aims to improve the transparency and robustness of AI models by using insights from neuroscience and statistics. He is also keen to apply AI to health and medical science, including areas such as drug discovery.

With a background in mathematics, Quentin earned an MSc in Data Science through a double-degree between ETH Zürich and CentraleSupélec. He completed his master’s thesis at the University of Oxford in computational neuroscience, under the supervision of Jascha Achterberg and Rui Ponte Costa. There, he developed a framework to better understand how AI models process information—essentially, how they turn inputs into outputs. This work complements the traditional approach of evaluating models only by their performance, by also looking inside the model to see how it represents and solves problems. His long-term goal is to embed abilities such as continual learning and flexible problem-solving into AI models, making them safer and more reliable.

Outside of research, Quentin completed three internships as an AI engineer in industries ranging from reinsurance to retail, giving him a broad perspective and practical expertise.