News

User Story

Engineering

Culture

David Wang

GPUs are often underutilized in production AI. Learn why CPU-centric architectures waste GPU capacity and how AI-native execution improves efficiency.

gpu inefficiency thumbnail 2x

GPU (In)efficiency in AI Workloads

KubeRay Team

KubeRay v1.5 introduces several enhancements for running Ray on Kubernetes, incorporating over 180 commits contributed by approximately 50 contributors.


Kuberay v1.5

Introducing KubeRay v1.5 

Masahiro Tanaka

Amjad Almahairi

Nithin Chalapathi

Matthew Deng

Praveen Gorthy

Richard Liaw

Richard Liaw headshot

Stephanie Wang

multimodal ai training thumbnail 2x

30% Faster Multimodal AI Training with Ray and Disaggregated Hybrid Parallelism

Seiji Eicher

Kourosh Hakhamaneshi

Rui Qiao

We are excited to announce new Ray Serve LLM APIs that make it easy to deploy state of the art serving patterns with vLLM

vLLM Ray Serve Anyscale - 3

Ray Serve LLM on Anyscale: Wide-EP and Disaggregated Serving with vLLM

Ray Serve LLM on Anyscale: APIs for Wide-EP and Disaggregated Serving with vLLM

Julian Forero

Heading to AWS re:Invent? Join us at booth #1854, book a 1:1 meeting or learn more across breakouts, lightning talks and executive round tables. 


anyscale re:invent 2025

AI with Ray and Anyscale at AWS re:Invent 2025

Emre Saglam

Every Ray server component now enforces authentication through a token validation middleware. This ensures that only requests with the correct token are processed by Ray’s internal and external services.


ray token based auth

Introducing Token-Based Authentication for Ray

Abrar Sheikh

Harshit Agarwal

Akshay Malik

Blog header - Ray Serve custom routing and more

Ray Serve: Advancing Flexibility with Async Inference, Custom Request Routing, and Custom Autoscaling

Alexey Kudinkin

Balaji Veeramani

ray data blog header image

Ray Data: Scalable Data Processing for AI workloads

Philip Wang

Philip Wang picture

Faster, cheaper and more resilient distributed AI processing with Anyscale Runtime powered by the Ray open-source framework

anyscale runtime blog header image

Announcing Anyscale Runtime for Faster, Cheaper and More Resilient AI, Powered by Ray

Justin Yu

Timothy Seah

Jason Li

Lehui Liu

Xinyuan Gui

ray train v2 blog header image

Ray Train V2: Unified Distributed Training on Ray

Christina Zhu

See all new Anyscale releases from Ray Summit 2025, including Anyscale on Azure preview, Anyscale Runtime performance updates, Global Resource Scheduler, multi resource cloud, and Lineage Tracking.

Roll up header image

Ray Summit 2025: Anyscale Platform Updates

Keerti Melkote, CEO of Anyscale

Ray has become the engine for AI-native computing for thousands of AI- and digital-natives. Now Anyscale and Microsoft are partnering to bring the power of Ray to every enterprise.

anyscale Microsoft partnership why now

Anyscale and Microsoft Partnership

Announcing first-party service - Anyscale on Azure

Announcing Anyscale Native Offering on Azure: Build, Run and Scale AI-Native Workloads Securely on Azure Infrastructure

Announcing Anyscale First-Party Offering on Azure: Build, Run and Scale AI-Native Workloads Securely on Azure Infrastructure

Ray, the open-source distributed compute framework for AI is joining the PyTorch Foundation, part of The Linux Foundation. 

Ray joins PyTorch Foundation

Ray is Joining The PyTorch Foundation

opensource stack blog thumbnail

Components of an Open Source AI Compute Tech Stack

An Open Source Stack for AI Compute: Kubernetes + Ray + PyTorch + vLLM

Explore a technical comparison of leading Reinforcement Learning (RL) libraries for LLMs from Ray. This guide analyzes frameworks like TRL, Verl, and RAGEN to help developers choose the best tools for RLHF, reasoning, and agentic AI.

The architecture of a Reinforcement Learning (RL) library is split into two primary components: Generation and Training. During the generation phase, an LLM Engine performs multi-turn rollouts within an environment to produce data and reward signals. This output is then fed into the training phase to update the model's parameters. This process forms a feedback loop, where the progressively improved model generates the next iteration of data for continuous refinement.

Open Source RL Libraries for LLMs Graph

Open Source RL Libraries for LLMs

This section is used to order the "Types" and "Tags" that show up for filters on the Blog Index

Types

Products / Libraries

Blog

Powered by Ray, Anyscale empowers AI builders to run and scale all ML and AI workloads on any cloud and on-prem.

Learn why Anyscale is best place to run Ray

Free, self-paced curriculum for practitioners

One-click to launch. Start fast, make it yours

Anyscale

Blogs