From Video to Voice: Build Faster with TensorRT Model Connect

TensorRT Model Connect makes it easy to go from open-source models to real, deployable applications without the usual integration overhead. In this demo, we will show two very different experiences built on the same TRT MC workflow: Cosmos-powered generation and a TRT Nemotron Voice full-duplex chat application. The goal is to show that TRT MC is not limited to one model category or one user experience. It gives developers a simple path to analyze models, build optimized runtimes, and deploy production-ready applications quickly, so teams can spend less time on deployment plumbing and more time building the product experience. What you'll learn: - How TRT MC simplifies the path from open-source models to real, deployable applications with a consistent workflow that is easy to use and easy to deploy. - How the same TRT MC workflow can support very different application experiences, from Cosmos-powered generation to a TRT Nemotron Voice full-duplex chat application. - How TRT MC reduces integration overhead by handling model analysis, optimized runtime generation, and deployment setup so developers can spend less time on plumbing and more time building the end-user experience. Join us live, bring your questions, and follow along as we show how TRT MC can be the foundation for building very different AI applications in real time.