Teodor Costăchioiu is a Research Scientist at the GeoSense Laboratory, Politehnica University of Bucharest. He joined GeoSense in December 2024, where he contributes to the AI4RISK project, developing AI-based methods for multi-source data fusion and predictive risk assessment.
His current work integrates Earth-observation data with advanced machine learning and retrieval-augmented generation (RAG) systems to support environmental monitoring and risk analysis.
He received his degree in Electronics and Telecommunications Engineering (a five-year program equivalent to M.Eng.) from Politehnica University of Bucharest in 1997, and his Ph.D. in Electronic Engineering from the same university in 2011. His doctoral research introduced a novel approach for spatio-temporal change detection using the Latent Dirichlet Allocation (LDA) algorithm, originally developed for text analysis, to model land-cover dynamics in satellite image time series of the Bucharest–Ilfov region.
Between 2011 and 2014, he was a Research Scientist at the CEOSpaceTech Research Center, where he contributed to the LEOSITS (Long-Term Data Exploitation for Satellite Image Time Series) project. His work during this period focused on the development of algorithms for spatiotemporal information extraction from large satellite archives.
His research interests include remote-sensing data analytics, AI-driven change detection, explainable machine learning, and the application of large language models for semantic interpretation of multimodal Earth-observation data.