
Welcome to my webpage!
I build AI systems for people who are usually missing from the data. When a crisis hits, the people hardest to see are the people hardest to help, and that is a measurement problem before it is a humanitarian one.
My work approaches it from two directions. One is infrastructure: mapping poverty from satellite and mobile data, forecasting cyclone impact before landfall, geolocating humanitarian reports, tracking displacement, and asking who actually receives an evacuation alert — and who doesn't. The other is our values: understanding what people believe, fear, and care about, because interventions fail when they collide with moral convictions. Both halves are the same question. What do people value, and what happens to the ones we fail to count?
Currently I am a Senior Research Scientist at UNICEF (New York) and at ISI Foundation (Turin), working across humanitarian AI and computational social science.
I hold a Ph.D. from the University of Trento, was a visiting Ph.D. researcher with the Human Dynamics group at the MIT Media Lab under the supervision of Sandy Pentland, and studied Electrical and Computer Engineering at the Technical University of Crete.
Selected papers, with BibTeX and code, are on my Publications page; the lexicons, models and datasets we have released are on Tools & Resources. The full record is on Google Scholar, DBLP and ORCID. Find me on LinkedIn, Bluesky and Twitter, or email me.
Evacuation Alerts and Socioeconomic Inequalities
#HumanitarianAI
Accessibility to Marketplaces in Africa
#HumanitarianAI
Toward a Global Impact-based Forecasting Model for Tropical Cyclones
#HumanitarianAI
Moral Expression in Different Domains
#MoralValues
Political Influence on Vaccination
#VaccineHesitancy
Geolocating Humanitarian Documents
#HumanitarianAI
Monitoring Ukrainian Immigration via LinkedIn Estimates
#SocialInequalities