Équipe de recherche :Research Team:
Cybersécurité et Cryptographie (C²)Cybersecurity and Cryptography (C²)
Laboratoire :Laboratory:
Laboratoire Traitement et Communication de l'Information (LTCI)Information Processing and Communication Laboratory (LTCI)
Département :Department:
Informatique et Réseaux (Infres)Computer Sciences and Networks (Infres)
Lê Nguyên Hoang has been an Assistant Professor at Télécom Paris, France, since 2026. He graduated from École Polytechnique (X07), France, and obtained his PhD degree from École Polytechnique de Montréal, Canada, with a best thesis award. He was then a postdoctoral researcher at MIT, USA, and EPFL, Lausanne, Switzerland. Later, he worked in science communication, entrepreneurship and associations, before joining Télécom Paris. He co-authored numerous books, including The Equation of Knowledge and Demain l’IA aux Commandes. He is also a member of the Orange AI Ethics committee.
His research focuses on AI governance and cybersecurity. Namely, how can learning systems be collaboratively constructed, with sensitive data protection, misuse resilience and influence traceability? He is especially interested in the extent to which recommendation and agentic AIs can be poisoned by malicious actors, and by the co-design of algorithms that comply with democratic norms.
Part of the research is incarnated in the nonprofit free and open-source Tournesol project that he co-founded. Tournesol provides a proof of concept of safe-by-design recommendation systems. The online Tournesol.app platform and Tournesol’s BlueSky feeds heavily rely on algorithms that emerged from his research.
His teaching focuses a thorough understanding of the cybersecurity vulnerabilities inherent to AI systems. It also presents the state-of-the-art risk mitigation techniques against privacy leakage, evasion attacks and data poisoning. Finally, it discusses real-time AI governance solutions.