Merritt Smith

Merritt Smith is a PhD student at the UC Berkeley School of Information, where he works to materially improve the lives of the world’s most impoverished people by increasing the effectiveness of humanitarian aid. His research combines machine learning and econometric techniques with both traditional data sources and mobile phone network data to better serve marginalized communities. He is particularly interested in leveraging existing relationship networks to enhance the targeting and scaling of interventions. He is grateful to be supported by the Hal Varian Fellowship and the World Bank Development Impact Group (DIME). 

Before joining the I School, Merritt worked as a data scientist at the University of Chicago Crime Lab and Center for Applied Artificial Intelligence, where he developed computer vision models to detect domestic violence in hospitals. He has also worked with the Centre for Economic Performance at LSE to give police evidence-based treatments to reduce domestic violence recidivism, built models to predict trade flows for the U.S. Census Bureau as a Civic Digital Fellow, and helped create population estimates for every block in sub-Saharan Africa at the Mansueto Institute. He holds an MS in Computational Analysis and Public Policy from the University of Chicago and a BA in Data Science and Public Policy from Tufts University.