Summary
Valentin Liévin is a Senior Research Scientist with a PhD and 11 years of experience at the intersection of large language models, information retrieval, and medical applications. He has led research and engineering efforts at DeepMind and Google to augment LLMs with search capabilities and improve LLM performance for healthcare and science. Valentin built retrieval-augmented systems and released the VodLM framework for training such models, and has developed faithfulness evaluation frameworks for LLM summaries. His academic work in deep generative models and variational inference underpins practical contributions in medical QA, NER, and deep retrievers for clinical retrieval. Comfortable moving between research and production, he combines rigorous probabilistic modelling with hands-on system delivery across startups and large labs. Based in Toronto, he brings a rare blend of theoretical depth and product-oriented engineering focused on trustworthy, retrieval-augmented NLP.
11 years of coding experience
7 years of employment as a software developer
Baccalauréat Sciences, Baccalauréat Sciences at Lycée baudelaire Roubaix
Mathématiques et physique (MPSI/MP), Mathématiques et physique (MPSI/MP) at Lycée Faidherbe - classe préparatoire aux grandes écoles
ingénieur généraliste spécialisation en Réalité Virtuelle, ingénieur généraliste spécialisation en Réalité Virtuelle at École centrale de Nantes
Technical University of Denmark
Master's degree Data interaction, Master's degree Data interaction at Korea Advanced Institute of Science and Technology