Saumya Sinha

Hello! I am a Postdoctoral Researcher at the National Laboratory of the Rockies (formerly named NREL) in Boulder, Colorado. My research lies at the intersection of advanced Machine Learning (ML) and the physical sciences, specifically within the domain of Scientific Machine Learning (SciML). I completed my Ph.D. in Computer Science at the University of Colorado Boulder in 2023, advised by Prof. Claire Monteleoni. My academic foundation was built at the Indian Institute of Technology (IIT) Kharagpur, where I earned my Integrated BS and MS in Mathematics and Computing.

At the lab, I am working in the Artificial Intelligence, Learning, and Intelligent Systems (ALIS) group, developing multimodal generative AI systems for building energy modeling and investigating LLMs to enhance scientific reasoning.

Previously, I have worked on modelling environmental impacts:

Sea-level Predictions: Building ML frameworks to fuse climate model simulations with satellite altimetry for future sea-level rise projections.
Renewable Energy Forecasting: Developing deep sequence learning architectures for probabilistic solar irradiance forecasting.
Extreme Event Detection: Using unsupervised deep learning to detect anomalies like avalanches from satellite imagery. (Spotlight talk at a NeurIPS workshop: link)

My long-term research goal is to develop physics-informed and interpretable AI systems that enhance scientific reasoning and transparency in high-stakes scientific applications.

Outside of research, I love to cook, watch tennis, and bike, and I am enjoying my time in Boulder!

Selected Publications

EnvDS

On the effectiveness of neural operators at zero-shot weather downscaling

Saumya Sinha, Brandon Benton, and Patrick Emami

Environmental Data Science 2025

Bib HTML

@article{sinha2025effectiveness,
  title = {On the effectiveness of neural operators at zero-shot weather downscaling},
  author = {Sinha, Saumya and Benton, Brandon and Emami, Patrick},
  journal = {Environmental Data Science},
  volume = {4},
  pages = {e21},
  year = {2025},
  publisher = {Cambridge University Press},
  note = {Discusses the trade-offs between Neural Operators and Transformers for super-resolution.}
}

E&B

Conditional Distribution Estimation of Building Characteristics with Diffusion Models for Urban Energy Modeling

Saumya Sinha, Alexandre Cortiella, Rawad El Kontar, and 3 more authors

Energy and Buildings Mar 2026

Bib HTML

@article{sinha2026conditional_eb,
  title = {Conditional Distribution Estimation of Building Characteristics with Diffusion Models for Urban Energy Modeling},
  author = {Sinha, Saumya and Cortiella, Alexandre and {El Kontar}, Rawad and Glaws, Andrew and King, Ryan N. and Emami, Patrick},
  journal = {Energy and Buildings},
  year = {2026},
  month = mar,
  doi = {10.1016/j.enbuild.2026.117392},
  note = {Published in Energy and Buildings. Applies conditional diffusion models to estimate building characteristic distributions for urban energy modeling.}
}

ICLR

SysCaps: Language Interfaces for Simulation Surrogates of Complex Systems

Patrick Emami, Zhaonan Li, Saumya Sinha, and 1 more author

In The Thirteenth International Conference on Learning Representations Mar 2025

Bib HTML

@inproceedings{emami2025syscaps,
  title = {SysCaps: Language Interfaces for Simulation Surrogates of Complex Systems},
  author = {Emami, Patrick and Li, Zhaonan and Sinha, Saumya and Nguyen, Truc},
  booktitle = {The Thirteenth International Conference on Learning Representations},
  year = {2025},
}

AIES

Multidecadal Sea Level Prediction Using Neural Networks and Spectral Clustering on Climate Model Large Ensembles and Satellite Altimeter Data

Saumya Sinha, John Fasullo, R Steven Nerem, and 1 more author

Artificial Intelligence for the Earth Systems Mar 2024

Bib HTML

@article{sinha2024multidecadal,
  title = {Multidecadal Sea Level Prediction Using Neural Networks and Spectral Clustering on Climate Model Large Ensembles and Satellite Altimeter Data},
  author = {Sinha, Saumya and Fasullo, John and Nerem, R Steven and Monteleoni, Claire},
  journal = {Artificial Intelligence for the Earth Systems},
  volume = {3},
  number = {4},
  pages = {e230089},
  year = {2024},
  publisher = {American Meteorological Society},
}

NeurIPS-W

Subseasonal Solar Power Forecasting via Deep Sequence Learning

Saumya Sinha, BM Hodge, and Claire Monteleoni

In NeurIPS 2021 Workshop on Tackling Climate Change with Machine Learning Mar 2021

Bib HTML

@inproceedings{sinha2021subseasonal,
  title = {Subseasonal Solar Power Forecasting via Deep Sequence Learning},
  author = {Sinha, Saumya and Hodge, BM and Monteleoni, Claire},
  booktitle = {NeurIPS 2021 Workshop on Tackling Climate Change with Machine Learning},
  year = {2021},
}

Variational autoencoder anomaly-detection of avalanche deposits in satellite SAR imagery

Saumya Sinha, S Giffard-Roisin, F Karbou, and 4 more authors

In Proceedings of the 10th International Conference on Climate Informatics Mar 2020

Bib HTML

@inproceedings{sinha2020variational,
  author = {Sinha, Saumya and Giffard-Roisin, S and Karbou, F and Deschatres, M and Karas, A and Eckert, N and others},
  booktitle = {Proceedings of the 10th International Conference on Climate Informatics},
  pages = {113--119},
  year = {2020},
  note = {Highly cited work establishing VAEs for cryospheric hazard detection.}
}