About

Research statement

My research integrates AI/ML, remote sensing, and geospatial modeling to address problems in soil, water, carbon, and agricultural systems across continental and farm scales. With a PhD in Computational Science grounded in optimization, decision-making under uncertainty, and the mathematical foundations of machine learning, I have applied this training as a Postdoctoral Research Scholar at Arizona State University to large-scale environmental data — developing reproducible frameworks for harmonizing legacy soil observations, quantifying continental-scale soil carbon storage potential, and benchmarking farm-scale water-use efficiency using OpenET, Sentinel-2 NDVI, and USDA crop and soil data. I also contribute to regime-aware multi-head LSTM architectures for operational streamflow forecasting in arid and semi-arid systems.

My future research program develops scalable, interpretable AI frameworks that integrate multi-source satellite, soil, climate, and hydrologic datasets to support monitoring, risk assessment, and management of agricultural and natural-resource systems.

Themes & methods

• Soil & carbon: continental-scale SOC/SIC modeling, dynamic baselines and attainable projections, depth-aware bias diagnostics, legacy-data harmonization.
• Water & agriculture: peer-conditioned attainable-ET benchmarks, NDVI-driven ET, crop-mask integration.
• Hydrology: ensemble streamflow forecasting with diverse loss functions, multi-head LSTM and Temporal Fusion Transformer architectures.
• Climate & cities: compound heatwave projection and urban resilience.
• Foundations: decision theory under uncertainty, interval computation, geometric/symmetry-based explanations of empirical regularities.

Tools

Python, R, MATLAB, C; PyTorch, TensorFlow/Keras, scikit-learn, XGBoost, CatBoost; Google Earth Engine, ArcGIS Pro, ENVI, GDAL, Rasterio, GeoPandas; Git, Docker, Jupyter; ASU Sol HPC cluster.

Languages

English (professional), Hindi (professional), Nepali (native).