Isaac Corley

Isaac Corley

Director of AI/ML Research

Taylor Geospatial

isaac.a.corley@gmail.com

I'm Isaac and I work on GeoAI — computer vision and foundation models for earth observation — from research through production. Currently Director of AI/ML Research at Taylor Geospatial, where I'm building the team and the models behind our earth observation work. Previously I built and shipped the RasterFlow platform at Wherobots for global-scale geospatial inference, and served as PI on the IARPA SMART program at BlackSky.

I maintain open-source projects — TorchGeo (4k+ stars), SMP (11k+ stars), and FTW — and publish country-scale and global-scale prediction maps as open data. I write a weekly blog with Caleb Robinson at geospatialml.com covering ML experiments for geospatial applications. Ph.D. in Electrical Engineering from UTSA.

News

Jul 2026

Presented TorchGeo-Bench at the EO-AI Symposium at TUM

Gave a talk introducing TorchGeo-Bench, our new evaluation harness for geospatial AI and foundation models, at the EO-AI Symposium hosted by TUM in Munich.

Jul 2026

New preprint: Fields of the Planet

We pair 3m PlanetScope imagery with Fields of The World annotations across 24 countries and show that higher-resolution imagery substantially improves field boundary delineation over 10m Sentinel-2.

Jun 2026

Presented at the CVPR 2026 Image Matching Workshop in Denver, CO

Our paper, Are Pretrained Image Matchers Good Enough for SAR–Optical Satellite Registration?, was presented at the Image Matching Workshop at CVPR 2026.

May 2026

Released the first global agricultural field boundary map at 10m resolution

Through Fields of The World, we released 3.17B field polygons across 241 countries as open data, with an accompanying preprint. Covered in the Taylor Geospatial announcement linked below.

May 2026

New preprint: No One Knows the State of the Art in Geospatial Foundation Models

Our 152-paper audit finds the GFM literature can't be ranked, with 46 cross-paper disagreements of 10+ points on the same model and benchmark. We propose six community standards to fix it.

Scroll for more

Projects

TorchGeo

TorchGeo

PyTorchGeospatialRemote Sensing

A PyTorch domain library, similar to torchvision, providing datasets, samplers, transforms, and pre-trained models specific to geospatial data.

Segmentation Models PyTorch

Segmentation Models PyTorch

PythonPyTorchSemantic Segmentation

A library containing a suite of PyTorch-based semantic segmentation decoders along with pretrained timm encoder support.

Fields of the World (FTW)

Fields of the World (FTW)

PythonPyTorchField Boundary Segmentation

A library for advancing machine learning models for instance segmentation of agricultural field boundaries in multispectral satellite imagery.

torchid

torchid

PythonPyTorchCUDA

GPU-accelerated intrinsic dimension estimators in PyTorch — a batched port of scikit-dimension with 12 estimators running end-to-end on CUDA tensors, achieving up to 2725× speedup over CPU baselines.

Selected Publications

arXiv preprint 2026

Fields of the Planet: Field Boundary Mapping Beyond 10m

Isaac Corley, Caleb Robinson, Jennifer Marcus, Hannah Kerner

arXiv preprint 2026

No One Knows the State of the Art in Geospatial Foundation Models

Isaac Corley, Nils Lehmann, Caleb Robinson, Gabriel Tseng, Anthony Fuller, Hamed Alemohammad, Evan Shelhamer, Jennifer Marcus, Hannah Kerner

arXiv preprint 2026

The First Global Agricultural Field Boundary Map at 10m Resolution

Caleb Robinson, Gedeon Muhawenayo, Subash Khanal, Zhanpei Fang, Isaac Corley, Ana M. Tárano, Lyndon Estes, Jennifer Marcus, Nathan Jacobs, Hannah Kerner, Inbal Becker-Reshef, Juan M. Lavista Ferres

CVPR Image Matching Workshop 2026

Are Pretrained Image Matchers Good Enough for SAR–Optical Satellite Registration?

Isaac Corley, Alex Stoken, Gabriele Berton

PRUE: A Practical Recipe for Field Boundary Segmentation at Scale

CVPR 2026

PRUE: A Practical Recipe for Field Boundary Segmentation at Scale

G. Muhawenayo, Caleb Robinson, S. Khanal, Z. Fang, Isaac Corley, A. Wollam, et al.

From Pixels to Patches: Pooling Strategies for Earth Embeddings

ICLR ML4RS 2026

From Pixels to Patches: Pooling Strategies for Earth Embeddings

Isaac Corley, Caleb Robinson, Inbal Becker-Reshef, Juan M. Lavista Ferres

Fields of The World: A Field Guide for Extracting Agricultural Field Boundaries

ICLR ML4RS 2026

Fields of The World: A Field Guide for Extracting Agricultural Field Boundaries

Isaac Corley, Hannah Kerner, Caleb Robinson, Jennifer Marcus

Earth Embeddings as Products: Taxonomy, Ecosystem, and Standardized Access

IGARSS 2026

Earth Embeddings as Products: Taxonomy, Ecosystem, and Standardized Access

Heng Fang, Adam J. Stewart, Isaac Corley, Xiao Xiang Zhu, Hossein Azizpour

HydroChronos: Forecasting Decades of Surface Water Change

ACM SIGSPATIAL 2025

🏆 Best Research Paper Candidate

HydroChronos: Forecasting Decades of Surface Water Change

Daniele Rege Cambrin, Eleonora Poeta, Eliana Pastor, Isaac Corley, Tania Cerquitelli, Elena Baralis, Paolo Garza

A Change Detection Reality Check

ICLR ML4RS 2024

A Change Detection Reality Check

Isaac Corley, Caleb Robinson, Anthony Ortiz

Barely-Visible Surface Crack Detection for Wind Turbine Sustainability

IROS 2024

🏆 Best Application Paper Runner-Up

Barely-Visible Surface Crack Detection for Wind Turbine Sustainability

Sourav Agrawal, Isaac Corley, Conor Wallace, Clovis Vaughn, Jonathan Lwowski

SSL4EO-L: Datasets and Foundation Models for Landsat Imagery

NeurIPS 2023

SSL4EO-L: Datasets and Foundation Models for Landsat Imagery

Adam J. Stewart, Nils Lehmann, Isaac A. Corley, Yi Wang, Yi-Chia Chang, Nassim Ait Ali Braham, Shradha Sehgal, Caleb Robinson, Arindam Banerjee

TorchGeo: Deep Learning with Geospatial Data

ACM SIGSPATIAL 2022

🏆 Best Paper Runner-Up

TorchGeo: Deep Learning with Geospatial Data

Adam J. Stewart, Caleb Robinson, Isaac A. Corley, Anthony Ortiz, Juan M. Lavista Ferres, Arindam Banerjee

Scroll for more

Full list on Google Scholar

Talks & Podcasts

Jul 2026

EO-AI Symposium, TUM

TorchGeo-Bench: An Evaluation Harness for Geospatial AI and Foundation Models

Invited talk introducing TorchGeo-Bench, our new evaluation harness for geospatial AI and foundation models, at the EO-AI Symposium 2026 in Munich (organized by Prof. Xiaoxiang Zhu's lab at TUM and the University of Queensland). Slides linked below.

Jun 2026

Satellite-Image-Deep-Learning Podcast

Mapping The World at Taylor Geospatial

Joined Robin Cole's podcast with Jennifer Marcus to discuss Fields of The World, openly licensed field boundary data, and geospatial foundation model benchmarking at Taylor Geospatial.

Watch recording

May 2026

Great Data Products Podcast

Fields of the World: Mapping Every Field on Earth

Joined Jed Sundwall's Great Data Products podcast with Jennifer Marcus to discuss the global release of Fields of The World, the fiboa schema, and what it takes to sustain open geospatial data products.

Watch recording

Apr 2026

UT Austin

Cloud Native GeoAI

Guest lecture at UT Austin on Cloud Native GeoAI, with slides and recording linked below.

Watch recording

Apr 2026

Clark University

Geospatial AI and Deep Learning with PyTorch

Guest lecture for Lyndon Estes' class at Clark University, with slides linked below.

Oct 2025

Spatial Stack Podcast

Beyond the Hype: Embeddings, Foundation Models, and the Future of Earth Observation

Joined Matt Forrest's Spatial Stack podcast with Chris Ren to discuss the current state of Geospatial Foundation Models and Embeddings.

Aug 2025

Satellite-Image-Deep-Learning Podcast

Chained Models for High-Res Aerial Solar Fault Detection

Joined Robin Cole's Satellite-Image-Deep-Learning podcast to discuss our CVPR PBVS paper: Aerial Infrared Health Monitoring of Solar Photovoltaic Farms at Scale.

Education

2020-2024
University of Texas at San Antonio

University of Texas at San Antonio

Ph.D. in Electrical Engineering

Advisor: Paul Rad

Thesis: Multimodal Learning for Mapping in Remote Sensing

2016-2018
University of Texas at San Antonio

University of Texas at San Antonio

M.S. in Electrical Engineering

Advisor: Yufei Huang

Thesis: Deep Learning for EEG Spatial Interpolation

2012-2016
Texas A&M University - Kingsville

Texas A&M University - Kingsville

B.S. in Electrical Engineering, Minor in Mathematics

Experience

2026 - Present
Taylor Geospatial

Director of AI/ML Research Taylor Geospatial

Building the AI/ML research team and shipping geospatial foundation models and earth observation pipelines from prototype through production.

2025 - 2026
Wherobots

Senior Machine Learning Engineer Wherobots

Built and scaled geospatial vision models powering the Wherobots spatial analytics platform. Shipped country-scale field boundary predictions and open-sourced prediction maps for 5 countries.

2021 - 2025
Zeitview (formerly DroneBase)

Senior Machine Learning Scientist Zeitview (formerly DroneBase)

Researched and deployed computer vision, vision-language, and 3D reconstruction models at scale for renewable energy inspection — solar farms, wind turbines, rooftops, and telecom infrastructure.

2024
Microsoft Research

Ph.D. Research Intern Microsoft Research

Advisor: Simone Fobi Nsutezo & Anthony Ortiz

Researched multimodal pretraining methods for large-scale geospatial vision-language datasets.

2021 - 2022
BlackSky

Senior Machine Learning Engineer BlackSky

Served as PI on the IARPA SMART program. Built and deployed models powering the Spectra AI platform's satellite image analytics.

2019 - 2020
HouseCanary

Senior Data Scientist HouseCanary

Built computer vision models extracting features from property images for HouseCanary's automated valuation model and property recommender.

2018 - 2019
Booz Allen Hamilton

Senior Data Scientist Booz Allen Hamilton

Prototyped deep learning methods for detecting image steganography and adversarially generated domains.

2016 - 2018
Southwest Research Institute (SwRI)

Research Engineer Southwest Research Institute (SwRI)

Advisor: Kenneth Holladay

Shipped software updates for the A-10 Warthog and researched ML methods for engine stall detection and MIL-STD-1553 bus exploitation.

2015
Oak Ridge National Laboratory (ORNL)

Research Intern Oak Ridge National Laboratory (ORNL)

Advisor: Paul Ewing

Collected and annotated a seismic dataset of human and vehicle activity and trained ML models to detect it.

© 2026 Isaac Corley