Martin Werner

Technical University of Munich (TUM) is advertising a PhD / doctoral candidate research associate position in Contrastive Learning and GeoAI within the Professorship Big Geospatial Data Management . The project focuses on hard negative sampling for contrastive representation learning and its applications in geospatial artificial intelligence . Research topics include developing sampling strategies that go beyond embedding similarity by integrating domain knowledge such as spatial distance, sensor metadata, and existing maps. A second part of the project explores how these strategies can identify informative subsets (coresets) of large geospatial datasets. Application domains mentioned include cross-view geo-localization and visual place recognition using aerial, street-view, and LiDAR data. The professorship emphasizes large-scale georeferenced data, distributed computing, machine learning, image and text analysis, randomized data structures, high-performance computing, and quantum algorithms. Eligibility highlights: a completed master’s degree in mathematics, computer science, physics, geoinformatics, data science, or related fields; strong machine learning background; very good programming skills (Python, C++, etc.); fluent English; German is a plus. The team values independent work, willingness to learn, and interest in an international research environment. Funding: full-time research associate position (100%, TV-L E13) for 3 years, funded by the German Research Foundation (DFG). The post includes the opportunity to pursue a doctoral degree. Application deadline: 1 August 2026. Start date is expected between September and November 2026. How to apply: prepare a single PDF with motivation letter, CV, degree certificates, transcripts, employment certificates, and any other relevant documents. Email the application to [email protected]. For questions, contact Prof. Dr. Martin Werner.