ray
@raydistributed
The AI framework trusted by OpenAI, Uber, and Airbnb. Created and developed by @anyscalecompute.
Huawei just released AsyncFlow, a new RL library for LLM post-training. arxiv.org/pdf/2507.01663
There have been a lot of open source RL libraries for training LLMs popping up recently. We took a stab at describing some of the use cases and design decisions they are optimized for: anyscale.com/blog/open-sour…
Interesting open source project aiming to simulate 10s of thousands of agents to model societal interactions (using @raydistributed). "Experiments demonstrate that the framework can support simulations of 30,000 agents that are faster than the wall-clock time."
DistFlow, from researchers at Shanghai Innovation Institute, looks like a cool new RL framework that builds on Ray, PyTorch, vLLM, and SGLang. arxiv.org/pdf/2507.13833 @raydistributed, @PyTorch, @vllm_project, @lmsysorg
This table was a footnote at the end of the blog, but it's actually one of the most interesting points. There is an emerging stack for post-training. anyscale.com/blog/ai-comput…
How @klaviyo uses Ray for data processing, training, hyperparameter tuning, and model serving! klaviyo.tech/ray-data-train…
Distributed computing is easy to start. Scaling it in production? Not so much. @raydistributed is the go-to for AI scaling — but running it at scale can get tricky. Anyscale makes Ray production-ready from the start. No hacks, no duct tape. 🗓️ Join us July 30 to learn: -Common…
Last year, the creators of @vllm_project at UC Berkeley hosted a massive two-day vLLM event featuring presentations from Roblox, Uber, Apple, Intel, Alibaba, Neural Magic, IBM, Handshake, Databricks, Anyscale, and others on how they are using and optimizing vLLM. This covered…
🚨Last call🚨 Tomorrow at 10am PT, join us live for AI in Practice: Build Scalable and Production-Ready MCP Servers. 4 live demos, real infra patterns, hands-on learning with Ray Serve and Anyscale 🙌 Register Now 👇 linkedin.com/events/7348418…
In large part due to Philipp's work on TRPO, reinforcement learning was one of the original motivating use cases that led us to build @raydistributed. You can see how we framed it in our early Ray paper (on page 1). arxiv.org/abs/1712.05889
Congratulations to my brilliant co-founder Philipp Moritz (@pcmoritz) and the legendary John Schulman, Sergey Levine, Pieter Abbeel, and Michael Jordan on their Test-of-Time Honorable Mention at ICML 2025 today! For creating TRPO. This was done during the previous wave of…
Ray will be at ICML!
🚀 Heading to #ICML2025 in Vancouver? Join us for drinks, snacks, & hot takes with Anyscale + @turingcom. - Kimi K2 - context engineering - lesson-based learning - the fossil fuel of AI 📅 Wed Jul 16, 7-10 PM downtown. Spots limited — RSVP 👉 lu.ma/gj4ir6fb
🚨Last chance to submit your CFP for Ray Summit 2025🚨 The Call for Proposals closes tonight. This is your final opportunity to speak at Ray Summit, happening in San Francisco, November 3–5. Have you: • Built something cool with Ray? • Solved a scaling challenge? •…
How WeChat (Tencent) Built #AI Infrastructure Supporting Millions of Nodes with Ray and Kubernetes (@raydistributed @kubernetesio) 👉 anyscale.com/blog/tencent-w…
Kimi K2 looks awesome. This blog post talks a little bit about how they use @raydistributed for data processing. alibabacloud.com/blog/kimi-larg…
🚀 Hello, Kimi K2! Open-Source Agentic Model! 🔹 1T total / 32B active MoE model 🔹 SOTA on SWE Bench Verified, Tau2 & AceBench among open models 🔹Strong in coding and agentic tasks 🐤 Multimodal & thought-mode not supported for now With Kimi K2, advanced agentic intelligence…
Scaling agentic systems? Part 2 of our Agentic AI and MCP series is for you. On July 23, we’re hosting a hands-on lab with four live demos on deploying MCP with Ray Serve and Anyscale – covering real-world patterns for production-ready systems. Save you spot:…
The Ray Summit CfP is closing on July 14! We'll be back in San Francisco from Nov 3-5 for Ray Summit and want to showcase your work. Whether it's scaling smoother, building GenAI workflows, or creating complex ML systems - if you've built it with Ray, we want to hear about it.…
🚨Last call!🚨 LLMs are evolving — your integration strategy should too. MCP (aka the “USB-C of AI apps”) fixes fragile wrappers, glue code & logging chaos. Join us tomorrow, July 9 at 10am PT for a live webinar: Unlocking AI with MCP: • How MCP evolved from LLM hacks to…
