Y Hacker News new | ask | show | jobs

user: robertnishihara created: 2017-10-16 karma: 462 submissions: Major upgrades to Ray Serve: 88% lower latency and 11.1x higher throughput 2 points | 1 comments SkyRL brings Tinker to your GPUs (2025) 24 points | 5 comments vLLM large scale serving: DeepSeek 2.2k tok/s/h200 with wide-ep 147 points | 54 comments Massively Parallel Agentic Simulations with Ray 2 points | 0 comments Deploy DeepSeek‑R1 with VLLM and Ray Serve on Kubernetes 1 points | 0 comments An Open Source Stack for AI Compute: Kubernetes and Ray and PyTorch and VLLM 1 points | 0 comments Native LLM APIs in Ray Data and Ray Serve 2 points | 0 comments Joins and Hash-Shuffle in Ray Data 3 points | 0 comments AsyncFlow: An Asynchronous Streaming RL Framework for LLM Post-Training 4 points | 0 comments Open Source RL Libraries for LLMs 1 points | 0 comments Large-Scale Deployment of Ray in Tencent's Weixin AI Infrastructure 2 points | 0 comments Uv and Ray: Pain-Free Python Dependencies in Clusters 44 points | 10 comments Roll: Reinforcement Learning Optimization for Large-Scale Learning 1 points | 0 comments An Open Source Stack for AI Compute: Kubernetes and Ray and PyTorch and VLLM 1 points | 0 comments Uv and Ray: Pain-Free Python Dependencies in Clusters 1 points | 0 comments Ray Batch Inference at Pinterest (Part 3) 1 points | 0 comments Direct Preference Optimization with Synthetic Data on Anyscale 1 points | 0 comments 0 points | 0 comments 0 points | 0 comments 0 points | 0 comments 0 points | 0 comments 0 points | 0 comments 0 points | 0 comments 0 points | 0 comments Building an LLM Router for High-Quality and Cost-Effective Responses 1 points | 0 comments Ray Infrastructure at Pinterest 1 points | 0 comments Lessons from training a Stable Diffusion model on 2B images 5 points | 0 comments Canva Built a Modern AI Platform Using Anyscale 2 points | 0 comments Building RAG-Based LLM Applications for Production 2 points | 0 comments