Steve BrumbyCo Founder

United States - Washington, District of Columbia

I am the Co-Founder and CEO/CTO of Impact Observatory, a mission-driven technology company bringing AI-powered geospatial monitoring to environment, climate, and sustainability risk analysis. Impact Observatory produced the world’s first fully automated, high resolution (10m) land use and land cover annual maps using deep learning at global scale in commercial cloud, released as a digital public good via our partners Esri Living Atlas and Microsoft Azure Planetary Computer. Impact Observatory's maps are also available via the UN Biodiversity Lab (powered by Impact Observatory), and have been used by the New York Times to analyze the impact of catastrophic flooding in Pakistan due to the climate crisis. Steve currently serves on the Department of the Interior's Landsat Advisory Group.

At National Geographic Society, I built and led a team to advise exploration and conservation programs, applying machine learning and scientific modeling to remotely sensed imagery and signals (satellites, aircraft, and drones), in situ networks (animal trackers, autonomous cameras and other sensors) and citizen explorer contributed datasets. We created and visualized these datasets to further the Society's mission of achieving a planet in balance through data-driven conservation, including supporting the largest scientific mission to take climate sensors to the top of Mount Everest in 2019.

I was the technical co-founder for Descartes Labs, a venture-backed company focused on understanding human activity and natural resources at global scale in real time through scientific datasets. Descartes Labs initial products serve agricultural, financial, insurance, and government/policy sectors with automated imagery analysis and business intelligence services. Descartes Labs was spun out of Los Alamos National Laboratory (LANL) in December 2014, where I was a Senior Research Scientist and Principal Investigator for LANL's Video Analysis & Search Technology (VAST), a neuroscience-inspired deep-learning algorithm for labeling social media video at scale. I led a team of 16 scientists applying deep learning to social media, surveillance video, satellite imagery and microscopy (biological and man-made materials). VAST ran on LANL's High Performance Computing systems and we were the first team at LANL to extend our work to commercial cloud systems (Amazon AWS). Prior to my deep learning research, I was co-inventor of a genetic algorithm system for image analysis called GENIE, that won an R&D Magazine R&D100 award in 2002, used for remote sensing and digital pathology applications.