Matteo Silvestri

I’m a second-year PhD student at Sapienza University of Rome.
I am a member of the HERCOLE Lab research laboratory, and I am currently a research trainee at the European Molecular Biology Laboratory.

My research focuses on using large language models to solve useful downstream tasks efficiently. For instance, I work on XAI Narratives, where LLMs are used to turn technical explainability outputs into human-friendly explanations while preserving faithfulness to the underlying model behavior. More broadly, I am interested in AI agents, especially in how they can leverage context to become more efficient, autonomous, and reliable.