Research / Publications
From sparse observations to continuous fields. Explore my work on scientific machine learning, surrogate models, and digital twins.
Google Scholar
An observation-driven digital twin that turns sparse stream-gauge histories, terrain, and rainfall into basin-wide flood predictions.
.png)
Reconstructing continuous spatiotemporal fields from sparse sensors with a temporal encoder and an implicit neural decoder.
.png)
Tensor-based reduced-order models that preserve spatial structure and compress unstructured mesh data for efficient flow-field prediction.
A region-optimal Gaussian process surrogate that partitions physical fields to accelerate cold-flow and combustion emulation.