I am a PhD student in Information Technology at Politecnico di Milano, working at NECSTLab on multi-agent systems for autonomous observability and self-healing cloud ecosystems. The question I keep returning to is how far a group of language-model agents can be trusted to diagnose a failing microservice on their own, and what has to be true of the evidence they are handed before their answer is worth acting on.
Research
Research
Publications and the questions behind them, pulled straight from ORCID.
Publications
1 published · 0 citations
2026Journal Article
Weakly Supervised Occupancy Prediction for HVAC Optimization in Digital Twins Smart Campuses
Marcello Martini, Fabio Maresca, Gurkan Solmaz, Flavio Cirillo, Luis Roda-Sanchez, Tobias Jacobs, Andrea Conti
IEEE Internet of Things Journal
0 citations
Refreshed from ORCID on 2026-09-28
Working on something adjacent?
The useful conversations start with a disagreement about the evidence.
If you work on agentic diagnosis, observability, or anything that has to justify an automated decision, I would rather hear where you think this is wrong than be told it sounds promising. A quarter of an hour is usually enough to find out whether there is something here worth doing together.