Geospatial ML · Open Source
TorchGeo
A PyTorch domain library for geospatial data — datasets, samplers, transforms, and pretrained models that make Earth-observation deep learning as easy as torchvision.
Anthony Ortiz
Principal Research Science Manager · Microsoft AI for Good Lab
I lead and conduct research at the intersection of machine learning, computer vision, and remote sensing, with a focus on applying AI to problems that matter for people and the planet — from geospatial machine learning and Earth observation to medical imaging and humanitarian response.
- Geospatial ML & Earth Observation
- AI for Climate & Sustainability
- Conservation & Biodiversity
- Medical Imaging
- Disaster & Humanitarian Response
Research
My work sits at the intersection of machine learning, computer vision, and remote sensing, with deep collaborations across health, sustainability, and humanitarian partners. I co-lead Microsoft’s Geospatial Machine Learning Center, and also do fundamental research on generalization, transfer learning, conditional computation, and domain adaptation.
Selected Work
Projects I’m proud of. See Research for more and Publications for the full list.
Health · 2025
AI-enabled screening for Retinopathy of Prematurity in low-resource settings
Deep-learning system to triage ROP from neonatal fundus images, validated in clinical settings where pediatric ophthalmologists are scarce.
Earth Observation · 2022
An AI inventory of every utility-scale solar farm in India
National-scale dataset built from satellite imagery to inform renewable-energy policy and siting decisions.
Earth Observation · 2021
Tracking glacial melt at the top of the world
Open data and ML to monitor the Hindu Kush Himalaya glaciers — work funded by an $18K Microsoft AI Research Grant.
Conservation · 2022
Detecting beached whales (and elk and cattle) from space
Very-high-resolution satellite imagery + deep learning for biodiversity and wildlife monitoring at scale.
Core ML · CVPR 2020
Local Context Normalization
A drop-in normalization layer that improves generalization of dense-prediction networks across geographies.
Selected Publications
Hand-picked highlights. Full list with year-grouped filters on Publications.
- TorchGeo: Deep Learning with Geospatial Data Community-standard PyTorch library for Earth observation; 159+ citations.
- AI-Enabled Screening for Retinopathy of Prematurity in Low-Resource Settings First-author clinical validation; covered by Ophthalmology Times.
- Global Renewables Watch: A Temporal Dataset of Solar and Wind Energy from Satellite Imagery Living atlas of global renewables; featured in The New York Times.
- An Artificial Intelligence Dataset for Solar Energy Locations in India First-author national-scale dataset informing renewable-energy policy.
- Effective Deep Learning Approaches for Predicting COVID-19 Outcomes from Chest CT Volumes First-author Nature-family publication; pandemic-response medical imaging.
- Local Context Normalization: Revisiting Local Normalization First-author CVPR; drop-in normalization layer for dense prediction.
- Mining Self-Similarity: Label Super-Resolution with Epitomic Representations Label efficiency for dense prediction at scale.
- Human-Machine Collaboration for Fast Land Cover Mapping Interactive ML system that turned weeks of mapping into minutes.
News
- 2025New paper: TorchGeo: Deep Learning with Geospatial Data in ACM TSAS.
- 2024Honored with Microsoft CELA’s Data Scientist of the Year Award.
- Mar 2024Promoted to Principal Research Science Manager at the Microsoft AI for Good Lab.
- Feb 2024Preprint out: A Change Detection Reality Check — a sobering benchmark for remote-sensing change detection.
- 2023Paper accepted in Computers in Biology and Medicine: weighted batch-wise dice loss for prostate-cancer metastasis segmentation in PSMA PET/CT.
- Sept 2021Promoted to Senior Research Scientist at Microsoft AI for Good.
Experience
-
March 2024 — Present
Principal Research Science Manager
Leading research programs in geospatial machine learning, medical imaging, and computer vision in support of Microsoft’s AI for Good initiatives.
-
Sept 2021 — March 2024
Senior Research Scientist
Microsoft AI for Good Lab
Led projects in geospatial ML, medical imaging, and computer vision in support of Microsoft’s AI for Earth and AI for Health programs.
-
June 2020 — Aug 2021
Research Scientist
Microsoft AI for Good Lab
Tackling large-scale problems in medical imaging, geospatial ML, and computer vision in support of Microsoft’s AI for Earth and AI for Health initiatives.
-
June — Aug 2019
Computer Vision / Deep Learning Research Intern
Conditional image generation with GANs for synthetic satellite imagery.
-
March — June 2019
AI Research Intern
Model generalization, human-machine collaboration, transfer learning, and land-cover mapping with Nebojsa Jojic and Dan Morris.
-
Aug 2018 — March 2019
Visiting Researcher
Mila · University of Montréal
Generalization, variational inference, and remote sensing for social good with Yoshua Bengio.
-
Aug 2017 — June 2020
Ph.D. Research Associate
Intelligent Agents and Strategic Reasoning Lab, UTEP
Conditional computation, adversarial ML, generative modeling and geospatial analytics with Christopher Kiekintveld.
Education
Ph.D., Computer Science
University of Texas at El Paso (UTEP) · 2015–2020
B.S., Telematics Engineering
Pontificia Universidad Católica Madre y Maestra (PUCMM) · 2010–2014
Mentoring
Selected interns and students I have had the privilege to mentor — current roles in parentheses.
- Isaac Corley · Research Scientist, WeRobotics
- Bibek Aryal · Software Engineer, Microsoft
- Alonso Granados · PhD Student, University of Arizona
- Zachary Bell · Senior Software Engineer, USAA
- Gerardo Uranga · Software Engineer, Walt Disney (prev. Microsoft, DocuSign)
Contact
anthony.ortiz [at] microsoft [dot] com
Elsewhere
Affiliation
Microsoft AI for Good Lab
Redmond, WA, USA