Blog

“MAS-Lab: An Open Framework for Spec-Driven, Interoperable Multi-Agent Systems” has been accepted at AGNTCon + MCPCon Europe 2026. I’ll be presenting this work and also attending as chair of the AAIF Accuracy and Reliability working group.

Interview with colleagues at the AAIF on open source and agentic AI. We discuss why open source matters for agentic systems, and the standards we’re building around agent failure modes, multi-agent observability, and operational practices for production AI systems.

We have released MAS-Lab, a specification-driven validation framework for reliable multi-agent systems, as open source on GitHub, together with an arXiv preprint.

Our paper on Aether, a framework combining multi-agent AI with network digital twins for autonomous network validation, is now available on arXiv.

Public Cisco Live EMEA 2026 talk video on agentic observability and evaluation, relevant to both network-validation and reliable-agentic-AI threads.

Elected chair of the Accuracy and Reliability working group in the AAIF Agentic AI Foundation, alongside Casper Nielsen from Diagrid as co-chair.

Technical whitepaper documenting our agentic AI platform’s production deployment results at Swisscom with detailed use cases.

Presentation on multi-agent systems for network change validation using knowledge graphs and AI orchestration.

A recommended talk by Giovanna Carofiglio on building reliable agentic AI — not my own work, but highly relevant to anyone thinking about trustworthy agent systems.

Presentation on agentic network change validation at Cisco Live US, covering AI-driven verification and NetDevOps automation.