Experience#
AI Special Interest Group (SIG) Leader
2026 - Present
M3AAWG
The Messaging, Malware and Mobile Anti-Abuse Working Group (M3AAWG) is where the industry comes together to work against bots, malware, spam, viruses, DoS attacks and other online exploitation. In parallel with my role at Hornetsecurity, I lead its Artificial Intelligence SIG:- Providing a forum for members to share insights on the evolving role of AI in enabling and combating abuse.
- Identifying gaps where collaboration or initiatives can improve outcomes, and contributing expertise to initiatives supported by Priority Committees.
- Focus areas: AI-generated phishing and scams, synthetic identities, automated content and abuse at scale, defensive applications of AI.
Lead Research Engineer
2022 - Present
Hornetsecurity (formerly Vade), a Proofpoint company
I lead research projects that bring Machine Learning, NLP and Computer Vision into production email protection, from the first experiment to the deployed detector.- Leading an NLP project that detects unsolicited textual emails, such as spear-phishing, Business Email Compromise and cold emails.
- Supervising two PhD students in partnership with public research laboratories, and publishing our work at venues such as ACL, EMNLP, LREC and CODASPY:
- Gabriel Loiseau: Privacy and Utility-Preserving Transfer Learning for Text Anonymization.
- Senaid Popovic: Intent detection in textual content.
- Presenting our research to the anti-abuse industry at M3AAWG General Meetings (LLM security, AI agents, LLM optimization).
- Continuing to mentor engineers, review code and shape engineering practices across the team.
Research Engineer
2017 - 2022
Vade
Vade, a French cybersecurity firm specializing in email security solutions, where I served as a Research Engineer within the Innovation team. Key responsibilities included:- Developing advanced technologies in C and Golang, leveraging Machine Learning, Computer Vision, and Natural Language Processing:
- Creation of PDF and Office document parsers, to extract meaningful features for Vade's malware classifier.
- Design of Computer Vision-based threat detection systems such as RIANA, successfully blocking up to 30 million emails monthly based on appearance or embedded remote images.
- Implementation of NLP-based detection mechanisms to identify textual threats such as spear-phishing.
- Providing technical mentorship, conducting code reviews, and establishing robust engineering practices.
- Delivering internal and external technical presentations.
- Participating in the recruitment efforts by conducting candidate reviews and interviews, overseeing internships.
Software developer
2015 - 2017
Stormshield
Stormshield, a leading French cybersecurity firm, is specialized in network, workstation, and data protection solutions.- Enhanced existing network analysis modules within custom intrusion prevention kernel module.
- Led integration of Gzip into HTTP module, conducting stress tests and benchmarks.
- Led the development of the High Availability prototype at kernel level.
PhD Student
2011 - 2015
2XS Team, CRIStAL, University of Lille
Network security study focused on Cloud Computing, using a passive approach.- Proposal of a distributed architecture of intrusion detection systems and a dedicated language to describe a set of security rules.
- Functional solution implemented in Python and C (~20,000 lines of code) including an intrusion detection system, a compiler and several modules.
- Position poster presented at Eurosys 2012, Bern, Switzerland.
- Publication of three peer-reviewed articles.
- PhD Thesis: Discus : Une architecture de détection d’intrusions réseau distribuée basée sur un langage dédié
Teaching assistant
2011 - 2015
University of Lille
- 370 teaching hours (BSc in Computer Science).
- Two main themes: System programming (memory allocation, process management, signals, debugging) and Networking (OSI layer, TCP/UDP protocols, routing and network equipment configuration).
- Supervision of internships and student projects.
Education#
I earned my MSc in Computer Science from the University of Lille in 2011, followed by the completion of a PhD in Computer Science in 2015, with a focus on cybersecurity.