Postdoctoral Associate

About the Role

The MUSE (Multimodal AI for Earth System Discovery) research cluster at Rice University is seeking a Postdoctoral Research Associate to pioneer multi-domain Graph Neural Networks (GNNs) and Foundation Models for Earth system modeling, state estimation, and forecasting. This position focuses on developing reusable, scale-adaptive, graph-based foundation architectures capable of learning unified representations across diverse Earth science testbeds. The goal is to develop transferable, physics-informed models that infer hidden Earth states, discover governing relationships, and quantify predictive uncertainties.

Initial term: One year, renewable for up to two to three years, subject to performance and funding.

Workplace: On campus in Houston, Texas.

Funding: Funded by a grant, soft and/or restricted funds. Continued employment is contingent on funding renewal.

Key Responsibilities

• GNN & Foundation Model Architecture Development: Design, build, and benchmark multi-domain graph neural network architectures and neural surrogates for complex spatial networks, irregular geometries, and spatio-temporal prediction tasks.
• Cross-Domain Transfer Learning: Integrate multimodal, heterogeneous Earth observations (e.g., seismic networks, river graphs, satellite altimetry, Distributed Acoustic Sensing) to evaluate model transferability across distinct Earth system domains (Water, Floods, Cryosphere, Earthquakes).
• Physics-Aware Machine Learning & UQ: Incorporate physical constraints (e.g., PDEs, mass/momentum conservation laws) into graph architectures and uncertainty quantification (UQ) frameworks.
• Proposal Development & Cluster Collaboration: Assist faculty with interdisciplinary research proposals (e.g., NSF, DOE, NASA) and collaborate with researchers across Computer Science, Electrical & Computer Engineering, and Earth, Environmental and Planetary Sciences (EEPS) at Rice.
• Dissemination: Publish research in top-tier machine learning venues (e.g., NeurIPS, ICML, ICLR) and geoscientific journals (e.g., AGU, Nature Geoscience) and present at national and international conferences.

Requirements & Tech Stack

• Python
• Deep learning frameworks: PyTorch, PyTorch Geometric / DGL
• GNNs
• Foundation Models
• Physics-informed models
• Uncertainty Quantification (UQ)
• High-performance computing (HPC)
• Cloud computing
• Distributed training (e.g., CUDA, multi-GPU training)

Qualifications

• Ph.D. in Computer Science, Data Science, Geophysics / Earth Sciences, Applied Mathematics, Electrical Engineering, Environmental Engineering, or a related quantitative field.
• Strong background in deep learning on graphs (GNNs), geometric deep learning, or large-scale foundation model architectures, with an interest or background in applying these methods to physical or geoscientific systems.
• High proficiency in Python and deep learning frameworks (e.g., PyTorch, PyTorch Geometric / DGL).
• Demonstrated track record of peer-reviewed publications in high-impact computational or geoscience journals and conferences.

Preferred Qualifications:

• Experience working with spatial-temporal or irregularly sampled physical data (e.g., remote sensing, streamflow networks, seismic arrays, InSAR, terrain data).
• Experience with Physics-Informed Neural Networks (PINNs), neural operators (e.g., FNO, DeepONet), or neural surrogate modeling for physical processes.
• Background in Uncertainty Quantification (UQ) or hybrid machine learning frameworks combining data-driven models with physical priors.
• High-performance computing (HPC) and/or cloud computing with expertise in distributed training (e.g., CUDA, multi-GPU training).

What We Offer

• Interdisciplinary Mentorship: Co-mentorship from a collaborative faculty team spanning ML, GNN theory, remote sensing, and geophysics.
• World-Class Computing: Full access to Rice’s High-Performance Computing (HPC) clusters and specialized GPU acceleration nodes.
• Competitive Compensation: Competitive salary and full Rice University benefits package.
• Career Development: Dedicated mentorship and career guidance tailored toward pathways in academia or industry research labs.

To Apply

Submit the following application materials via the Rice University Application Portal as a single PDF file:

• Cover Letter (highlighting relevant research experience with GNNs, foundation models, or Earth science data)
• Curriculum Vitae (CV)
• Research Statement (1–2 pages detailing research accomplishments and future interests)
• Contact Information for 3 Professional References

Application review begins on December 1st, 2026, and continues until the position is filled.

Informal inquiries can be sent to Rice EEPS (esci-postdocs@rice.edu) or cluster faculty members.