My research develops AI-driven disaster geoinformatics by integrating photogrammetry, remote sensing, GIS, and explainable artificial intelligence. I aim to transform diverse observations—including satellite, SAR, UAV, LiDAR, DEM/DSM, field surveys, and infrastructure data—into reliable and actionable knowledge for disaster risk reduction. Disaster-related data are often incomplete, asynchronous, and acquired at different spatial scales. They may also be affected by cloud cover, shadows, occlusions, and limited ground observations. To address these challenges, I embed photogrammetric geometry, three-dimensional structure, terrain continuity, and hazard-process knowledge into AI models. This enables robust and uncertainty-aware analysis of floods, landslides, urban damage, infrastructure disruption, and post-disaster change. Beyond conventional static maps, my goal is to create dynamic four-dimensional representations that integrate space, time, semantics, and uncertainty. Such representations can support rapid situation assessment, field-survey prioritization, recovery monitoring, and evidence-based resilience planning. Through collaboration with IRIDeS researchers, local governments, and international partners, I will develop shared datasets, practical analytical workflows, and educational opportunities that connect advanced Earth observation technologies with safer and more resilient societies.