Actionable Repair via Multimodal AI
This exploratory workshop investigates how multimodal AI could engage with repair practices for a sustainable built environment. Using a Language-to-Action workflow, participants analyse damaged artifacts by integrating text, image, and geometric data. Unstructured inputs are organized into coherent spatial representations of the artifact and its damage condition. Participants query a LLM (Gemini 3.0) to propose repair strategies through zero-shot or single-shot prompting. These strategies are manually enacted, supported by AI guidance via a mobile, web-based interface. By bridging digital analysis with physical action, the workshop explores how AI can support creative reuse and extend material lifespans through informed, human-led repair.
Tutors
Kathrin Dörfler (TU Munich, Munich, Germany)
Kathrin Dörfler is Associate Professor of Digital Fabrication at the Department of Architecture at the Technical University of Munich. She holds a PhD from ETH Zurich and a Master‘s from TU Vienna. Her work focuses on computational design, on-site robotics, and collaborative fabrication.
Tizian Rein (TU Munich, Munich, Germany)
Tizian Rein is an architect and researcher bridging computational design and building preservation. During his Master‘s studies, he gained research experience in robotics at TU Munich and ETH Zurich. He is currently a doctoral researcher at the Professorship of Digital Fabrication within the Munich Institute for Robotics and Machine Intelligence (MIRMI) and MIT, USA.
Begüm Saral (TU Munich, Munich, Germany)
Begüm Saral is an architect and researcher specializing in mobile robotics and AI for construction and deconstruction. A graduate of ITU and TUM, she is currently a PhD candidate at TUM GNI. Her research within the "SPAICR" project focuses on spatial AI for robotic material reclamation and reuse.