PhD Candidate · Civil & Environmental Engineering · Cornell University
My work sits at the intersection of AI, hydrology, and climate risk. I develop physical, machine-learning, and physics-informed ML models to study environmental systems, quantify risk, and improve predictions.
View Research
Hi! I'm Sandeep, a PhD student in Civil and Environmental Engineering at Cornell University and a Graduate Research Assistant with the Steinschneider Research Group.
I grew up in Nepal and came to the United States for graduate school. My work sits at the intersection of AI, statistics, and hydrology. I study water, environmental, and climate systems and aim to help improve their resilience in a changing climate.
When I'm not working with data and models, I'm usually somewhere out in nature, reading a sci-fi novel, watching a good show, or playing a few games of chess.
Cornell University · Ithaca, NY
University of Connecticut · Storrs, CT
Institute of Engineering, Pulchowk · Kathmandu, Nepal
Steinschneider Research Group · Cornell University · Ithaca, NY
Knighton Ecohydrology Lab · University of Connecticut · Storrs, CT
Hydro-Consult Engineering · Nepal
Cloud Factory · Nepal
Browse my research projects on GitHub.
Integrating neural networks with process-based hydrological models in a end-to-end gradient-based framework to improve out-of-sample prediction and fidelity under climate change.
Evaluating attention-enhanced LSTM networks against standard architectures to improve accuracy in ungauged basin prediction.
A Bayesian hierarchical framework for regionally pooling flood-change projections from deep learning and hybrid models, reducing uncertainty for water infrastructure planning.
Investigating uncertainty in climate-driven design flood projections across process-based, deep learning, and hybrid hydrological models under varying levels of precipitation uncertainty.
Investigating the impact of future climate risks to US housing market and National Flood Insurance Program.
Analyzing historical housing data from NY/NJ/CT, finding that NFIP participation, risk perception, and flood memory are key regional drivers of home values following Hurricane Sandy.
I'm always happy to connect about research, collaboration, or just a conversation.