Anyscale

As more teams scale open source AI on secure, governed infrastructure, Ray has become critical across the full AI lifecycle, from data curation to production serving and reinforcement learning. Ahead of PyTorch Conference North America 2026 in San Jose, we put together a guide highlighting key Ray-focused sessions on the schedule. Featured speakers include: Anyscale / Databricks / University of California, Berkeley: Ion Stoica Anyscale: Eric Tang, Sumanth R Hegde, Joshua Lee, Mengjin Yan Google: Ankita Luthra, Kotturu Trinadh, Jago Macleod LinkedIn: Tao Huang, Tommy Li Pinterest: Gaurav Arora, Shunyao Li, Eric W. Uber: Ke Chen, Peng Zhang, Xandra Zhu University of Minnesota: Arun Sharma Explore engineering takeaways from these technical sessions where you will learn about unified AI orchestration with Kubernetes, elastic training stacks, and scalable RL. Read the full guide and map out your schedule: https://lnkd.in/e373G-Hu Register for PyTorch Conference North America: https://hubs.la/Q04tBgv_0 #PyTorchCon

As more teams scale open source AI on secure, governed infrastructure, Ray has become critical across the full AI lifecycle, from data curation to production serving and reinforcement learning. Ahead of PyTorch Conference North America 2026 in San Jose, we put together a guide highlighting key Ray-focused sessions on the schedule. Featured speakers include: Anyscale / Databricks / University of California, Berkeley: Ion Stoica Anyscale: Eric Tang, Sumanth R Hegde, Joshua Lee, Mengjin Yan Google: Ankita Luthra, Kotturu Trinadh, Jago Macleod LinkedIn: Tao Huang, Tommy Li Pinterest: Gaurav Arora, Shunyao Li, Eric W. Uber: Ke Chen, Peng Zhang, Xandra Zhu University of Minnesota: Arun Sharma Explore engineering takeaways from these technical sessions where you will learn about unified AI orchestration with Kubernetes, elastic training stacks, and scalable RL. Read the full guide and map out your schedule: https://lnkd.in/e373G-Hu Register for PyTorch Conference North America: https://hubs.la/Q04tBgv_0 #PyTorchCon

We’re heading to CoreWeave's Fully Connected 2026 next week! Meet the Anyscale team at Booth #600 and catch Xinyu Zhang on Thursday, October 1 at 4 PM on his session on: Data Pipeline at Petabyte Scale: Production Video Curation with Ray on CoreWeave Xinyu will share how a unified architecture can eliminate data silos and streamline video curation at petabyte scale, along with benchmarks demonstrating gains in throughput, utilization, TCO, and more. If you’re attending #FullyConnected2026, come say hello! 👋 https://lnkd.in/gCVQ3Cwu

I watched a bunch of the talks at Ray Summit live and this one by Priunsh Syen and Tyler Titsworth at Lila Sciences about the details of their AI research platform was extremely impressive. https://lnkd.in/eJxyTCwp

Ray Summit 2026, in two minutes. Three days in San Francisco. Keynotes on where Ray and Anyscale are headed. Breakouts from Netflix, Pinterest, Spotify, Zoox and more. The vLLM Conference. And the community that builds Ray. Every recording is now live on our YouTube channel: keynotes, breakouts, vLLM sessions, and Day 0 training. Watch on demand: https://lnkd.in/g5ZdsBX9

Ray Summit 2026 is a wrap. More than 2,000 people came to San Francisco on August 24-26 to ask the question: what does it actually take to run reinforcement learning at scale? Lila Sciences is turning a 15,000-square-foot robotics lab into a verifier for scientific RL, with results spanning CAR-T development, mRNA design, and alloy screening.  Torc Robotics consolidated five training systems onto one Ray cluster and is now running one billion autonomous trucking simulation miles per week.  Bryan Catanzaro from NVIDIA walked through how Ray coordinates 3,000-plus GPUs for Nemotron RL training, with a 13 percent throughput improvement from topology-aware scheduling alone.  Periodic Labs, Bedrock Robotics, Spotify, Capital One, Microsoft AI, Recursion, and more brought the same depth to the breakouts. Read the full recap. https://lnkd.in/gwrGc-zH

That's a wrap on Ray Summit 2026. Three days in San Francisco with over 2,000 engineers, researchers, and AI leaders working on the four frontiers: Foundation Model Training, Multimodal Data Curation, Physical AI, and LLM Reinforcement Learning. → Keynotes from Anyscale CEO Keerti Melkote and co-founder Robert Nishihara, with customer stories from Lila Sciences, Torc Robotics, and the vLLM community → 90+ breakouts and lightning talks from the teams running Ray in production → A full Training Day of hands-on workshops, and the first-ever vLLM Conference → New announcements across Ray and Anyscale, from KubeRay Connect to the Ray history server Thank you to every speaker, attendee, and partner who made it happen. Missed the keynote? Watch the replays. Day 1: https://lnkd.in/gdq_Z7ba Day 2: https://lnkd.in/gTWwuBVE

Learned today that Periodic Labs uses Kimi to reduce inference costs by 20-50x over frontier APIs. Everyone will do this. Inspirational talk by Liam Fedus.

Ray passed 170 million downloads in a single quarter. That's 5x growth in one year. The keynote at Ray Summit 2026 covered what's behind that growth: organizations are moving from using models to building their own, and closing the loop so those models improve continuously. → Keerti Melkote on the shift from consuming AI to building AI, joined on stage by Nscale CEO Josh Payne → Robert Nishihara on what it actually takes to run a learning loop, from RL orchestration to weight sync → Customer keynotes from Lila Sciences CTO Andrew Beam, Torc Robotics Head of AI Felix Heide, and Inferact co-founder and CEO Simon Mo, lead maintainer of vLLM Read more on the path to owning your intelligence: https://lnkd.in/gNSeKBXh

Announced at Ray Summit: Anyscale GPU Health Observability A GPU fault has always looked identical to a broken script from the outside. Not anymore. → DCGM signals (XID errors, ECC counts, SM Clock, per GPU memory) enriched with the exact job and workspace running on that GPU, automatically → One correlated view instead of five disconnected surfaces, drill straight from a failed job to the GPU behind it → Works across KubeRay and VM deployments today, with K8s Anyscale Operator support coming soon Learn more: https://lnkd.in/gdQxamPN