Research

Prithvi-WxC-powered, scalable, cost-efficient, high-resolution ensemble precipitation forecasting for flood-risk assessment and operational decision support.

Machine-learning-based flash flood forecasting in West Africa using satellite observations, numerical weather prediction, and distributed hydrologic modeling.

Machine-learning-based discrimination of seismic event types using curated waveform data, geophysics-inspired feature transformations, and configurable decision pipelines.

Topography-assisted downscaling of coarse water fraction maps to high-resolution flood inundation maps via deep learning approaches.

Bottom-up, at-scale simulation of the spread and evolution of information cascades in online social media.