Dr. Philipe Dias

Philipe Dias is an R&D Associate in Machine Learning and Computer Vision at Oak Ridge National Laboratory, where he designs AI-driven workflows for large-scale satellite imagery analysis. His recent work focuses on developing and pretraining multimodal foundation models for geospatial imagery, advancing generative AI for geospatial data forecasting and synthesis, while leveraging high-performance computing for training and deploying computer vision systems. His research spans the full image analysis pipeline, including image annotation, active learning, and evaluation efforts for applications such as building footprint extraction, disaster damage assessment, and land-use characterization. He has also contributed to the research community as chair and co-organizer of international workshops and tutorials at the CVPR and WACV conferences. Prior to joining ORNL, his work found application in wastewater treatment, agricultural automation, and healthcare-related scenarios. Philipe received a Double Master’s Degree in Information Technology from Hochschule Mannheim (Germany) and in Electrical & Computer Engineering from the Federal University of Technology (Brazil), and earned his Ph.D. in Electrical and Computer Engineering from Marquette University (USA).