Dr. William Keely is a Postdoctoral Fellow at NASA’s Jet Propulsion Laboratory in Pasadena, California. His work centers on applying generative, probabilistic, and physics-informed machine learning methods to inverse problems in remote sensing. William’s research includes diffusion-based retrievals of atmospheric trace gases for EMIT, OCO-2 and OCO-3, real-time onboard temperature mapping for AVIRIS-3 to support wildfire response and asset management, and novel Bayesian optimization methods for data quality filtering and correction. William is passionate about infusing machine learning with robust uncertainty quantification to produce trustworthy scientific products.

Prior to his current role, William worked as an intern at JPL with both the Machine Learning & Instrument Autonomy (MLIA) group and the Tropospheric Composition Group. He also worked as a Data Scientist in industry, bringing machine learning methods into operational Earth observation workflows, and served as a Research Assistant for NASA’s GeoCarb mission, developing novel retrieval algorithms.

Find my publications on Google Scholar

Contact

Education

• Machine Learning (PhD) - University of Oklahoma
• Mathematics (BS) - University of Oklahoma

Research Interests

• Diffusion Generative Modeling
• Inverse Problems
• Imaging Spectroscopy
• Physics/Forward Model Emulation
• Uncertainty Quantification
• SciML