I am an Artificial Intelligence researcher with expertise in foundation models, computational decision-making, and multi-agent systems. My work has been applied to diverse domains, including global-scale geospatial intelligence, autonomous mobility networks, and robotics.
At Google Research, I develop foundation models and applications for geospatial intelligence, human mobility, and the built environment. My contributions have driven interdisciplinary published research and real-world product impact across Google Maps, Ads, and other key platforms. Previously, as Research Lead at a Series A startup, I built intelligent digital platforms that optimized urban transportation and advanced smarter cities.
I hold a Ph.D. in Computer Science from Stanford, where my award-winning research on multi-agent decision-making for large-scale autonomous systems was covered by the BBC, Venture Beat, and IEEE Spectrum. I also earned an MS in Robotics from Carnegie Mellon, focusing on advanced robot planning techniques.
Recent News
- May 2026: Our paper Mobility-Embedded PoIs was accepted for publication at ICML 2026! ME-PoIs learns multimodal multi-task representations of points of interest by augment text-only embeddings with large-scale human mobility signals.
- March 2026: S2Vec, our paper on global-scale self-supervised geospatial embeddings, was published in the ACM Transactions on Spatial Algorithms and Systems journal and highlighted by the Google Research Blog.
- October 2025: Our vision track paper on Mobility Foundation Model Embeddings was published at ACM SIGSPATIAL 2025. Congratulations to my collaborator and student researcher, Maria Siampou, who gave a brilliant spotlight talk!
- July 2025: We're organizing the first ever workshop on Urban Mobility Foundation Models (UMFM) at ACM SIGSPATIAL 2025! Please check out the workshop website for details.
- April 2025: Multiple showcases of my mobility foundation model work on the Google Research blog: Geospatial Reasoning and Mobility AI!
Selected Publications
Mobility-Embedded POIs: Learning What A Place Is and How It Is Used from Human Movement
International Conference on Machine Learning (ICML), 2026.
Paper
S2Vec: Self-supervised geospatial embeddings for the built environment
ACM Transactions on Spatial Algorithms and Systems (TSAS), 2026.
Paper
