AI Operator: Foundation models for zero-shot industrial bin picking
Jimmy Li - Vention
Jan. 23, 2026, 2:30 p.m. - Jan. 23, 2026, 3:30 p.m.
ENGMD 279
Hosted by: Hsiu-Chin Lin
This talk presents a robotic system for unstructured industrial bin picking developed at Vention, a Montreal-based industrial automation company. We highlight how modern vision foundation models (FoundationStereo, Segment Anything 3, and FoundationPose) enable the rapid development of a zero-shot perception pipeline, achieving exceptional robustness and versatility without any training. To validate this approach, we showcase the system in action for a commercial client application. We demonstrate the ability to pick a variety of plumbing parts from a bin and insert them into a receptacle that feeds a manufacturing assembly line, proving the pipeline's readiness for real-world deployment.
Jimmy Li is the Director of Physical AI at Vention, where he leads a team focused on making vision-guided bin picking solutions accessible to all manufacturers. He spearheaded the development of Vention’s AI perception pipeline for robotic bin picking by leveraging state-of-the-art vision foundation models and played a pivotal role in strengthening Vention's partnership with Nvidia to accelerate the commercialization of this technology. Prior to Vention, Jimmy was a researcher at the Samsung AI Center in Montreal, working on reinforcement learning for 5G network optimization. He holds a PhD in Computer Science from McGill University, advised by Gregory Dudek and David Meger, where his thesis focused on using semantic objects as landmarks for simultaneous localization and mapping.
