I build computational tools to understand how people move, consume, and impact the environment. Here is a selection of my work bridging AI, systems, and climate science.
Do we really need invasive location data to understand mobility? We proposed a privacy-preserving Transformer model that uses only speed sequences to detect transportation modes, achieving SOTA accuracy (95.97%) and robust cross-regional generalization.
Climate change isn't just about warming—it's about volatility. We discovered a 14% increase in the day-to-day variability of the Southern Hemisphere's dominant climate mode, driven by thermodynamic shifts that fuel more extreme weather events.
Academic models often stay in the lab. I led the engineering of a production-grade system that processes terabytes of messy GPS data to track real-time carbon footprints, serving 300+ users and bridging the gap between research algorithms and reality.