RAY
FSDP and PyTorch Enable Large-Scale Model Training
Fully Sharded Data Parallel (FSDP) in PyTorch, integrated with Ray, optimizes GPU memory usage for scalable training of models like Qwen3-TTS with 1.7B parameters.
Adyen (ADYEN) Trains AI Model on 51 Trillion Tokens, Tackling Fraud
Adyen unveils its Transaction Foundation Model, trained on 51 trillion tokens, aiming to enhance fraud detection and payment optimization.
Anyscale Launches Debugging Skills to Streamline Ray and vLLM Fixes
Anyscale's new debugging tools simplify fixing Ray and vLLM workloads, saving hours of manual effort for developers.
Ray Day NYC Spotlights AI Scaling from Coinbase, Discord, Torc
Ray Day NYC featured Coinbase, Discord, and Torc Robotics sharing how Ray boosted AI workloads. Highlights include Torc's 90% GPU utilization.
Anyscale Launches Persistent Ray Dashboards for Debugging AI Workloads
Anyscale introduces new Cluster and Actor dashboards for Ray, offering full data persistence and enhanced debugging for distributed AI workloads.
Anyscale Launches Agent Skills to Streamline AI on Ray
Anyscale's new Agent Skills enhance AI coding tools like Claude Code and Cursor, optimizing Ray-based workflows for speed and scalability.
Notion Slashes AI Embedding Costs 80% After Ditching Spark for Ray
Notion migrated from Spark on EMR to Ray, cutting embedding costs 80% and improving query latency 10x. Uber and Salesforce shared similar AI infrastructure wins.
Ray 2.55 Adds Fault Tolerance for Large-Scale AI Model Deployments
Anyscale's Ray Serve LLM update enables DP group fault tolerance for vLLM WideEP deployments, reducing downtime risk for distributed AI inference systems.
VLA Models Reshape Robotics as $94B Market Embraces AI Infrastructure
Vision-Language-Action models are driving robotics teams to Ray and Anyscale for distributed training. Market projected to hit $94.38B by 2031.
Ray's Disaggregated Hybrid Parallelism Boosts Multimodal AI Training by 30%
Ray's innovative disaggregated hybrid parallelism significantly enhances multimodal AI training efficiency, achieving up to 1.37x throughput improvement and overcoming memory challenges.
Anyscale Showcases AI Innovations at AWS re:Invent 2025
Anyscale highlights AI solutions with Ray at AWS re:Invent 2025, featuring demos, talks, and executive events for enhanced machine learning operations.
Ray Enhances Scheduling with New Label Selectors
Ray introduces label selectors, enhancing scheduling capabilities for developers, allowing more precise workload placement on nodes. The feature is a collaboration with Google Kubernetes Engine.
Enhancing Ray Clusters with NVIDIA KAI Scheduler for Optimized Workload Management
NVIDIA's KAI Scheduler integrates with KubeRay, enabling advanced scheduling features for Ray clusters, optimizing resource allocation and workload prioritization.
Enhancing Text-to-SQL Models Using Tinker and Ray
Discover how Tinker and Ray are utilized to fine-tune text-to-SQL models, enhancing AI capabilities in generating efficient SQL queries.
Tencent's Weixin Integrates Ray for Large-Scale AI Deployment
Tencent's Weixin team has embraced Ray and Kubernetes to enhance their AI infrastructure, tackling challenges in resource utilization and deployment complexity.