He conducts research on the characterization of extreme rainfall and its implications for hydrological risk assessment. His work examines the characteristic of extreme rainfall events and utilizes radar-based precipitation data to overcome the spatial limitations of conventional rain gauge observations.

​

He also explores rainfall estimation from imagery using convolutional neural networks, aiming to extend instantaneous estimates to longer accumulation periods. By integrating these approaches with hydrological and statistical modelling, he seeks to improve hydrological risk assessment and support the design of resilient water infrastructure.