Anyscale

Ray Day: Seattle recap is here. Notion: 80% cost cut on embeddings with Ray on Anyscale. Salesforce: 200K token docs summarized at P95 < 15s. Uber: 20% better GPU utilization with heterogeneous Ray clusters. Apple: foundation model training at scale. Full story: na2.hubs.ly/H04MTFW0

Next stop: NYC. Join us for Ray Day NYC on May 21 at Convene, One Liberty Plaza for practitioner-led talks and a hands-on workshop focused on scaling AI systems in production. Spots are limited (invite required): anyscale.com/ray-day-nyc

Ray on the Road kicked off in SF with engineers building real systems. Now going global. Deep dives on distributed training, LLM scaling, and production lessons. Next: Seattle (3/31), Toronto (4/14), Boston (5/19), NYC (5/21) Invite @ na2.hubs.ly/H04zkNH0

If you’re building or scaling AI systems, join @robertnishihara, co-founder of Anyscale and co-creator of Ray, for a panel sharing lessons from teams moving AI from proof-of-concept to production. "HX26_RNISHIHARA" to save $150 on your ticket: na2.hubs.ly/H04zF5Q0

Ray Serve introduces two improvements delivering up to 11x higher throughput and 88% lower latency • HAProxy replaces Python HTTP proxy • Direct gRPC between replicas reduces overhead na2.hubs.ly/H04vxjb0

🥰🥰

The RL training infra for Composer 2 was built with Ray. Very impressive work.

AI search at web scale 🌐 Join @lancedb, @anyscalecompute, and @ExaAILabs for a deep dive into the stack behind Exa’s search engine — built on Ray + Lance. 📍 AWS Builder Loft, SF 👉 na2.hubs.ly/H04nbJv0

Join robotics experts from @mimicrobotics, @multiplylabs, @nebiusai and Anyscale as they share lessons and experiences around: • Sourcing, scaling, and managing 1000s of GPUs • How to scale AI without linearly scaling DevOps 👉 na2.hubs.ly/H04m6ps0

Last call! Sign up for Tuesday's live, virtual hands-on lab where you’ll code alongside our lead instructor and get your questions answered in real time. Register: na2.hubs.ly/H0431YQ0

@bonsai_robotics is building next-gen off-road autonomy, starting in agriculture. As multimodal datasets and models scaled, on-prem GPUs became a bottleneck. With Anyscale they are running scalable AI in the cloud. na2.hubs.ly/H04cHmY0

3 days until Ray Workshop SF. Hands-on, 200-level deep dive into scaling AI workloads with Ray. Mar 10 | JW Marriott SF. Invite required: na2.hubs.ly/H03Ztfw0

@bonsai_robotics uses Anyscale to train large VLA models for outdoor autonomy, powering multi-ton hydraulic equipment in agriculture. na2.hubs.ly/H044rYR0

Attending NVIDIA GTC? Join the AI Developers Meetup to discuss advanced AI workflows, from reinforcement learning and agent evals to next-gen notebooks. Hosted with @CoreWeave, @wandb, @marimo_io, @OpenPipeAI, and @Monolith_AI. March 17 | San Jose Register: site.wandb.ai/resources/even…

Ray Workshop SF: • Real-world scaling patterns • Live, hands-on labs • Direct access to Ray experts • Peer networking with AI engineers Mar 10 | JW Marriott SF. Invite required: na2.hubs.ly/H03Yrs00

Last call! Sign up for tomorrow's webinar on how VLAs are transforming robotics and autonomous systems – enabling more adaptive, general-purpose automation while keeping proprietary data secure and governed. Register: na2.hubs.ly/H03YrKL0

Curious how Ray scales AI training in practice? We’re hosting a live, virtual hands-on lab where you’ll code alongside our lead instructor and get your questions answered in real time. Register: na2.hubs.ly/H040M9-0

Proud to share Anyscale was named to @Forbes America’s Best Startup Employers 2026. This recognition belongs to our incredible team behind the platform that powers AI at scale. na2.hubs.ly/H0426P70

Ray Workshop is coming to Toronto (Apr 14) at the Startwell - Main Campus. Hands-on, 200-level training to scale AI workloads with Ray (Core/Data/Train/Serve). Request invite: na2.hubs.ly/H03ZrlS0