Apr 04 2024 44 mins
An interview with Mani Golparvar, co-founder and CTO of Reconstruct, a software platform for construction project management. Mani discusses how Reconstruct uses AI and computer vision to generate 3D models from images and video captured on construction sites. This allows stakeholders to remotely monitor progress and compare as-built reality to design models. Benefits include reducing travel costs, improving project visibility, and verifying installed work. Mani also shares his experience as a professor researching AI for construction at the University of Illinois. He provides insights on construction vs design workflows for AI, challenges acquiring training data, and emerging techniques like synthetic data. Overall, the interview explores AI's transformative potential in construction through Mani's dual lenses as an academic researcher and technology entrepreneur. Chapters : 00:09 - Introduction00:17 - About Reconstruct01:00 - Using visual data in construction02:00 - Background 04:13 - Managing time as professor and startup founder 07:00 - Challenges faced 09:00 - From research to entrepreneurship11:00 - Rule-based vs data-driven AI techniques in construction16:00 - Using computer vision for construction safety23:00 - Short vs long term AI research needs in construction 24:00 - Getting industry data for AI research31:00 - Potential data licensing models 32:00 - Tapping into existing hardware and data38:00 - Receptiveness to robotics in construction39:00 - Multimodal models with schedule, images, 3D40:45 - Rapid fire questions#contech #aectech #computervision
--- Send in a voice message: https://podcasters.spotify.com/pod/show/mayur-m-mistry/message