About

A physicist by training, working where statistical methodology, machine learning and the physical sciences meet.

Background

The question has been the same since my undergraduate thesis: how do you establish, rigorously, that data contain something your model did not predict? I trained as a physicist at the University of Padova and carried out my doctorate at CERN, the European laboratory for particle physics, as a CERN Doctoral Student in the CMS experiment — one of the two general-purpose detectors at the Large Hadron Collider. I was advised by Marco Zanetti and Andrea Wulzer and supervised at CERN by Maurizio Pierini. My thesis developed the New Physics Learning Machine, a model-agnostic test for departures from the Standard Model, and confronted what I still consider the central obstacle to using such methods for real measurements: keeping a p-value trustworthy when the reference model itself carries systematic uncertainty. I deployed the method on CMS collision data and on detector monitoring.

From 2023 to 2026 I held an IAIFI Fellowship — a competitively awarded independent postdoctoral fellowship, which gave me the freedom to define my own research programme across MIT and Harvard. I used it to take anomaly detection beyond the setting it came from: into learned representations, into the validation of generative models, and into the behaviour of large AI models. Over those years I was mentored by Philip Harris at MIT and Demba Ba at Harvard SEAS, and built collaborations across particle physics, statistics and machine learning.

I am now a Postdoctoral Associate at the MIT Laboratory for Nuclear Science. I remain a member of the CMS collaboration, having convened its ML Innovation group from 2023 to 2025, and I serve on the organizing team and Advisory Committee of PHYSTAT, the international forum where particle physicists and statisticians work out how inference should be done in the field. I co-organize VERAIPHY, which is setting out the statistical standards I would like work in this field to be held to.

Education and appointments

Fellowships, funding and awards

Languages

Italian (native), English (fluent), Catalan (intermediate), French (basic).