Post-training Nemotron with Ray | Bryan Catanzaro (Nvidia) | Ray Summit 2026 @anyscale
Post-training Nemotron with Ray | Bryan Catanzaro (Nvidia) | Ray Summit 2026  @anyscale
Uploaded September 2026 | Updated November 2026, 4 days ago
"Demand for intelligence is effectively infinite, which means AI will always run at the limit of available compute. Once you accept that, efficiency is intelligence.

Bryan Catanzaro is VP of Applied Deep Learning Research at NVIDIA. His research led to the creation of cuDNN, the library beneath nearly every deep learning framework, and he helped lead the teams behind DLSS and Megatron. In this talk, he walks through how NVIDIA builds Nemotron, its open family of models, datasets, and techniques: pre-training in 4-bit arithmetic, hybrid SSM-transformer architectures, an ensemble of domain-expert teacher models distilled into one, and reinforcement learning across 15 environments, coordinated by Ray on thousands of Blackwell GPUs. He also explains why NVL72 exists and why open technologies matter for every company building its own AI.

Liked this video? Check out other Ray Summit keynote sessions:
youtube.com/playlist?list=PLYnBpswCtPo4

Subscribe to our YouTube channel to stay up-to-date on the future of AI: youtube.com/c/anyscale

🔗 Connect with us:
LinkedIn: linkedin.com/company/joinanyscale
X: https://x.com/anyscalecompute
Website: anyscale.com/"
Post-training Nemotron with Ray | Bryan Catanzaro (Nvidia) | Ray Summit 2026Apple’s Approach to Scalable Machine Learning Infrastructure on Ray | Ray Summit 2025Serving Fast and Smart: Getting the Most Out of Ray Serve LLM on GKE | Google | Ray Summit 2026Anyscale on Azure: Build, Train, and Serve AI in Your Own Tenant | Microsoft | Ray Summit 2026Improved Scheduling Flexibility with Label Selectors in Ray | Ray Summit 2025Meet verl: An RL Framework for LLM Reasoning & Tool Use | Ray Summit 2025Scaling Machine Learning at Discord: Our Journey with Ray and Anyscale | Discord | Ray Summit 2026How BMW Scales Automotive AI Workloads with the Ray Framework | Ray Summit 2025Scaling Frontier AI with Ray on Kubernetes | KubeRay | Ray Summit 2026Ray Train on Anyscale: Observability and Reliability at Scale | Anyscale | Ray Summit 2026State of vLLM 2025 | Ray Summit 2025Training AI on the Scientific Process | Liam Fedus (Periodic Labs) | Ray Summit 2026
Anyscale |

Post-training Nemotron with Ray | Bryan Catanzaro (Nvidia) | Ray Summit 2026

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER