NVIDIA Hosts Livestream on Building Visual AI Agents

NVIDIA is set to host a livestream on November 18, 2025, focused on building visual AI agents utilizing its Cosmos Reason and Metropolis platforms. This event aims to address the challenges AI systems face in connecting perception with reasoning within dynamic real-world environments.

According to the NVIDIA Developer Blog, the Cosmos Reason Visual Language Model, or VLM, integrates vision, language, and world knowledge to enhance intelligent video and multimodal understanding. During the livestream, participants will have the opportunity to learn how to post-train Cosmos Reason with their own datasets, which is crucial for tailoring AI agents to specific applications.

The session will include hands-on demonstrations and practical insights on creating intelligent workflows that can be applied across various sectors, including manufacturing, logistics, and safety. Sources indicate that the event will highlight real-world use cases, showcasing the potential for advanced robotics and automation solutions powered by NVIDIA's technology.

By leveraging NIM microservices and the VSS blueprint, attendees can gain a deeper understanding of how to construct AI agents capable of functioning effectively in complex environments. This initiative underscores NVIDIA's commitment to fostering innovation in robotics, particularly in how AI can complement and enhance robotic capabilities in industrial settings and beyond.

As industries increasingly turn towards automation, NVIDIA's livestream represents a significant step in equipping developers and engineers with the tools needed to advance their robotics applications.