Webinars

NVIDIA Cosmos and Embed in Production

Building physical AI is hard. Labeling the data for it shouldn't be. Egocentric robot video is some of the most complex data. Making sense of it often means many annotators, weeks of setup, and workflows that can't keep pace with your models. Join NVIDIA and Encord for a live session on how Cosmos and Embed are being put to work inside real annotation pipelines. What You'll Walk Away With: See Cosmos and Embed in action for automatically captioning, classifying, and describing egocentric robot video A repeatable pipeline for automated pre-labeling on robot video, deployable from day one How to use Cosmos Embed to find the edge cases without watching hours of footage A live demo of plugging world models into you workflows to ship robotics applications at scale & a peak into Cosmos Reason 3 Who This Is For: Robotics Engineers Perception Engineers Computer Vision & ML Engineers Applied AI Leads Your Hosts: Kai has built annotation infrastructure for some of the most frontier robotics teams on the planet. He knows where physical AI data pipelines break, why they break, and exactly how to fix them. Nick sits at the intersection of cutting-edge robotics development and real-world AI deployment, working directly with the teams building the next generation of humanoids, warehouse robots, and autonomous systems. Register to attend live or receive the recording.