Muhammad Talha

Profile

Dual-major Ph.D. candidate developing AI for Earth-system prediction through multimodal foundation models, graph neural networks, spatiotemporal transformers, and transferable representations. At Oak Ridge National Laboratory, contributing to the Genesis-funded HydroORBIT project and graph-based hydrologic modeling. Experience integrating environmental observations, geospatial attributes, and numerical-model simulations.

Experience

1. Aug 2026 – Present

   Research Intern, Graduate Research Opportunity (GRO) Program

   Oak Ridge National Laboratory · Oak Ridge, Tennessee

   Mentor: Dr. Dan Lu

   • Contributing to HydroORBIT, a Genesis-funded multimodal foundation model for continental-scale hydrologic prediction, integrating environmental observations and simulations.
   • Developing graph neural networks for distributed streamflow prediction using river-network connectivity, NWM pretraining, and gauge observations, with benchmarking against neural models and TVA forecasts.
   • Building aligned multimodal datasets from gridded fields, basin attributes, and gauge time series, including point-cloud representations for spatial and temporal synchronization.

2. Jan 2023 – Jul 2026

   Graduate Research Assistant

   Michigan State University · East Lansing, Michigan

   Advisors: Dr. Amirpouyan Nejadhashemi and Dr. Pang-Ning Tan

   • Developed HydroTransformer and an explainable watershed-agnostic CNN–LSTM for streamflow prediction from distributed inputs, evaluating generalization across space and time.
   • Developed Basin2Vec, a basin-centric representation-learning framework for transferable hydrologic prediction in previously unseen basins.
   • Created BASINGRID, a 40-year (1985–2024) dataset of aligned meteorological and landscape rasters covering 9,067 U.S. basins.

3. Jun 2021 – Dec 2022

   Machine Learning Engineer

   Ibex (TRG)

   • Developed a real-time compliance system processing over 20,000 agent screen recordings using CNN-based form detection and OCR-based information extraction.

Education

1. Jan 2023 – Present

   Ph.D., Computer Science & Biosystems and Agricultural Engineering (dual major)

   Michigan State University · East Lansing, Michigan

   • Dissertation: Spatially Distributed Basin Representations for Hydrologic Prediction Across Space and Time.
   • Advisors: Dr. Amirpouyan Nejadhashemi and Dr. Pang-Ning Tan.

2. Aug 2017 – Jun 2021

   B.S., Computer Science

   University of Engineering and Technology, Lahore · Lahore, Pakistan

Publications

1. Basin2Vec: Learning Basin-Centric Embeddings for Scalable Hydrological Modeling Muhammad Talha, Xiaofeng Liu, Yuan Yang, Pang-Ning Tan, A. Pouyan Nejadhashemi IEEE/CVF Winter Conference on Applications of Computer Vision (WACV 2027)
2. BASINGRID: A distributed 40-year spatiotemporal hydrology dataset for U.S. basins Muhammad Talha, A. Pouyan Nejadhashemi, Xiaofeng Liu, Yuan Yang Scientific Data, 2026
3. Watershed-agnostic modeling: An explainable, distributed, data-driven hydrologic model Muhammad Talha, A. Pouyan Nejadhashemi Environmental Modelling & Software, 2026
4. Soft computing paradigm for climate change adaptation and mitigation in Iran, Pakistan, and Turkey: A systematic review Muhammad Talha, A. Pouyan Nejadhashemi, Kieron Moller Heliyon, 2025
5. Unveiling the resilience of smallholder farmers in Senegal amidst extreme climate conditions Kieron Moller, A. Pouyan Nejadhashemi, Muhammad Talha Food and Energy Security, 2024
6. Transferable Learning Architecture for Scalable Visual Quality Inspection Muhammad Talha, Sheikh Faisal, Zain Iftikhar, M. Touseef Afzal International Conference on Artificial Intelligence (ICAI), 2022
7. Multimodal Foundation Priors for Diffusion-Calibrated Zero-Shot Time Series Reconstruction Yingda Fan, Muhammad Talha, Vinh Ngoc Tran, Chonghao Qiu, Xi Xiao, Xiaowei Jia, Dan Lu Under review, International Conference on Learning Representations (ICLR)
8. From Aggregated to Distributed: A Transformer Framework for Generalizable Streamflow Modeling Across Space and Time Muhammad Talha, A. Pouyan Nejadhashemi, Xiaofeng Liu Under review, Engineering Applications of Artificial Intelligence

Full list, including manuscripts under review

Additional research

1. Jun 2023 – Jan 2024

   Multi-objective Optimization for Smallholder Farmers

   Michigan State University. Applied linear and multi-objective optimization using SciPy and pymoo (NSGA-II), ranking solutions by cost-effectiveness and nutritional balance.

Awards and fellowships

• 2026 Graduate Research Opportunity (GRO) Program · Oak Ridge National Laboratory
• 2025 M. Kent Taylor Graduate Assistantship · Michigan State University

Technical expertise