Research
I have been affiliated with Cisco Systems France since 2014 and currently work within Outshift as a tech lead. My work and expertise sit at the intersection of systems, networking, security, and performance evaluation. Over the years I have worked on topics ranging from network measurement and information-centric networking to cloud threat detection and AI-driven operations, always trying to balance theoretical and practical contributions.
Since March 2026, I chair the Accuracy & Reliability working group within the Agentic AI Foundation (AAIF), where we work on bringing agentic applications to the higher standards of reliability required by demanding enterprise environments.
Current research
My current work focuses on reliable multi-agent AI systems — building platforms for observability, evaluation, and interpretability of agentic applications, alongside runtime and lab infrastructure for reliable design and reproducible experiments.
2025-present
Observability, evaluation, and interpretability platform
Ontology-grounded observability and analysis of agent execution traces (CLARIS ontology), with a revised metrics computation engine (MCEv2) for systematic evaluation of multi-agent system behavior.
2025-present
Agent runtime, controller, and experimentation lab
The MAS Framework (agent-runtime, mas-ctl, mas-lab) — a runtime and controller for agentic applications, with a lab for reliable design and reproducible experiments. Designed for open-source release.
Public project pages: Internet of Agents · Internet of Cognition
Publications & Patents
Peer-reviewed papers, journal articles, IETF drafts, patents, and presentations are on the publications page.
Previous research
Nov. 2024 – Oct. 2025
Cisco / Outshift
Agentic network change verification and network digital twins
AI-driven validation pipelines, multi-agent orchestration, and network digital twins for change verification and operational assessment.
Jun. 2023 – Oct. 2024
Cisco / Outshift
Cloud threat detections
Realtime threat detection through adversarial threat generation, realtime knowledge graph generation from system calls and cloud events, rule-based and AI-based detection. Contributions to the CDR engine of Panoptica's CNAPP.
2015 – 2021
Cisco Systems France
Information-Centric Networking
Architectures and systems questions around information-centric communication, including mobility and quality of experience for gaming and video conferencing.
2016 – Jun. 2021
Hybrid ICN
Applied research project around hICN and vICN, focused on deployable ICN over IP, experimentation support, and public implementations in the FD.io ecosystem.
Jul. 2019 – Apr. 2020
QoS and scheduling for ICN
Short-term project on QoS and scheduling, bridging the hICN and cloud-security periods.
Sept. 2014 – Feb. 2015
Telecom ParisTech
Active network measurements, analysis, and anycast census
Anycast deployment characterization, anycast studies, and large-scale distributed measurement campaigns using fastping at Telecom ParisTech.
2009 – 2014
UPMC / LIP6 / LINCS
Testbed federation and active network measurements
Distributed active measurements, routing analysis, measurement platforms (libparistraceroute, TopHat, MySlice), and testbed federation at UPMC/LIP6/LINCS.
2008 – 2009
University of Cambridge
Passive measurements
Passive measurement work on high-bandwidth links at the University of Cambridge, within the MASTS project. Poster (PDF).
2004 – 2008
Orange Labs / Telecom ParisTech
Flow-Aware Networking
PhD at Orange Labs, jointly with Telecom ParisTech. Flow-aware architectures, performance modelling, and predictable behavior in software routers and access networks.
Artefacts liés
2-
20260421 webpage mapme, hicn, icn, research
-
20260421 webpage irtf, anrp, award, research