avatar

Hi, I’m Léo

I am a postdoctoral researcher in cybersecurity at the SnT (Interdisciplinary Centre for Security, Reliability and Trust), University of Luxembourg. I received my Ph.D. from IMT Atlantique in 2024 and hold an engineering degree in information security from ENSIBS. Before my doctorate, I worked as an application-security engineer at Orange Cyberdefense.

My research focuses on reliable AI for collaborative cybersecurity: how organizations can share knowledge and learn together while preserving security, robustness, and interpretability.

Research

My Ph.D. investigated federated learning for collaborative intrusion detection in distributed systems. It addressed data heterogeneity, resistance to poisoning attacks, reputation-based assessment of participant contributions, and reproducible datasets for evaluating federated intrusion detection.

At SnT, I am the scientific and technical coordinator of COCTEL (Causality-Oriented Cybersecurity TELemetry), an FNR CORE project led by Jérôme François. COCTEL studies the monitoring and security of distributed systems, especially microservices and AI systems, through telemetry and provenance records. I also contribute to PAIGE (Platform for AI Governance and Evaluation), a Luxembourg Ministry of the Economy PRDI project with Thales Cyber Solutions Luxembourg. In PAIGE, I contribute to the cybersecurity-monitoring and governance dashboard and co-supervise work on AI-system monitoring.

My current interests include:

Selected publications

My record comprises 14 international publications: 2 journal articles, 5 international conference papers, 5 international workshop papers, and 2 ICDCS tutorial papers. I also have 2 peer-reviewed national conference papers.

  1. Léo Lavaur, Marc-Oliver Pahl, Yann Busnel, and Fabien Autrel. “The Evolution of Federated Learning-based Intrusion Detection and Mitigation: A Survey.” IEEE Transactions on Network and Service Management, special issue on Network Security Management, June 2022. The survey proposes a reference architecture and taxonomy for federated intrusion detection; it had 125 Google Scholar citations on 1 April 2026.
  2. Léo Lavaur, Yann Busnel, and Fabien Autrel. “Investigating the Impact of Label-flipping Attacks against Federated Learning for Collaborative Intrusion Detection.” Computers & Security, volume 156, article 104462, 2025. This journal article extends the systematic evaluation published at ARES/BASS 2024; together, the two papers had 39 Google Scholar citations on 1 April 2026. Its evaluation code and results are available through Eiffel and cose_2025.
  3. Léo Lavaur, Pierre-Marie Lechevalier, Yann Busnel, Romaric Ludinard, Géraldine Texier, and Marc-Oliver Pahl. “RADAR: Model Quality Assessment for Reputation-aware Collaborative Federated Learning.” 43rd International Symposium on Reliable Distributed Systems (SRDS), IEEE, September 2024. RADAR combines model clustering, cross-evaluation, and a reputation system to identify unreliable contributions in collaborative federated learning. It had 10 Google Scholar citations on 1 April 2026; the implementation is available on GitHub.

My complete publication record is available on Google Scholar.

Teaching

I have taught more than 120 hours in undergraduate, engineering, and master’s programs in France and Luxembourg. My teaching includes network security (IPsec, 802.1X, DNSSEC, and TLS), networking, IoT security, intrusion detection and security monitoring, and agentic AI. At the University of Luxembourg, I designed and delivered the 9-hour 2025–2026 Agentic AI lecture course and contribute 11 hours of lectures and practical work in 2026–2027. I also co-authored and delivered federated-learning and network-security tutorials at NoF 2023 and ICDCS in 2024 and 2025.

My teaching work includes designing lectures, practical sessions, and self-contained Python notebooks. I have also mentored engineering apprentices at ENSIBS, acting as a link between students, their employers, and their workplace supervisors.

Academic service

My service includes co-chairing the program and organizing committees for CNSM MAGIC 2026 and NeSecOr 2025; serving on the program committees for ESORICS ANUBIS 2026 and NOMS MCT 2026; and organizing RESSI 2026 and Shadow PC AlgoTel-Cores 2026–2027. I have reviewed for IEEE TNSM, Computer Networks, JNSM, ACM Computing Surveys, IEEE Internet of Things Journal, IEEE/IFIP NOMS, IEEE CNSM, and SRDS.

I previously co-chaired the doctoral track program committee for RESSI 2024 and serve as chair of the “Distributed Learning and Security” session at WiMob 2026. Through the Shadow PC initiative, I help early-career researchers gain reviewing experience without participating in paper selection.

Recognition