PhD in AI for Earth Observation (Space AI)

About the Role We are looking for exceptional PhD candidates to join the AI for Earth Observation (Space AI) group , working on the next generation of Space AI systems — from foundation models for Earth observation to agentic geospatial pipelines deployed at the frontier of research. As a PhD student, you will work on cutting-edge topics including multi-modal fusion of satellite, aerial, and ground-level data, spatiotemporal reasoning, generative geospatial intelligence, and agentic Earth observation systems. You will contribute to state-of-the-art systems such as FireScope, our wildfire risk prediction model presented at CVPR 2026, and work with massive geospatial datasets spanning a variety of modalities and applications. You will be mentored by leading experts — Dr. Danda Pani Paudel and Prof. Luc Van Gool — and collaborate with top AI labs worldwide. PhD candidates are eligible for the prestigious Google DeepMind PhD Fellowship. The position is fully funded for up to five years with an outstanding fellowship of approx. €40.000 gross per year. Responsibilities Conduct world-class research in Space AI: foundation models for Earth observation, multi-modal fusion, spatiotemporal reasoning, and generative geospatial intelligence. Publish in top academic venues (CVPR, ICCV, NeurIPS, AAAI, and similar). Develop large visual-language models that reason over satellite, aerial, and ground-level imagery, text, spatiotemporal signals, and geospatial context. Build systems producing decision-ready outputs: reports, risk rasters, dense segmentation maps, and vector geospatial layers for applications such as disaster assessment, infrastructure monitoring, ecology, and geoscience. Design agentic Earth observation pipelines that autonomously select data, chain analytical steps, and verify results with minimal human supervision. Work with massive real-world geospatial datasets, addressing missing data, domain shifts, and physical/geometric priors. Collaborate with other PhD students, supervise Master's students, and engage with partner labs internationally. Present research progress clearly in English, INSAIT's working language. Required Qualifications Bachelor's or Master's degree (completed or nearing completion) in Computer Science, Data Science, Mathematics, Physics, Statistics, or Electrical Engineering. Strong academic record with high-quality coursework and grades. Solid programming skills, including experience with deep learning frameworks such as PyTorch. Genuine motivation for research and a passion for creating impact on a global scale. Strong command of written and spoken English. Preferred Qualifications Master's degree in AI, ML, Computer Vision, or a closely related field (Bachelor's-only applicants complete a Master's during the first two years of the program, fully funded). Research experience in computer vision, multimodal learning, remote sensing, or geospatial AI. Publications or research projects in relevant areas. Experience with large-scale training, foundation models, or vision-language models. Familiarity with Earth observation data: multispectral/hyperspectral imagery, SAR, or geospatial toolchains. Formal English proficiency certificate (TOEFL, IELTS, Cambridge, or similar). What We Offer Full PhD fellowship of approx. €40.000 gross per year for up to 5 years. Mentorship by world-class researchers and eligibility for the Google DeepMind PhD Fellowship. Collaboration and fully covered research visits with top international AI labs. Official degree issued by INSAIT at Sofia University. Rolling admissions — apply at any time. Ideal Candidate Profile A highly motivated researcher-in-the-making who wants to push the frontier of AI for understanding our planet. Curious, independent, and technically strong — excited by hard problems that span multimodal learning, spatiotemporal reasoning, and real-world deployment, and driven to produce research that matters at a global scale. International applicants are warmly welcomed; all work and communication at INSAIT takes place in English. Learn more about our Space AI research: vision.insait.ai/research/topics/space-ai