Catch up on keynotes, technical sessions, and announcements from Microsoft Build 2026.
OVERVIEW
A managed Ray platform for ML engineers, AI developers, and platform architects
- From multimodal data curation and model training to production deployment – build unified AI applications across data processing, embedding, training, and inference using a single, unified Python runtime. Anyscale on Azure provides a fully managed Ray runtime on AKS with automated cluster provisioning, autoscaling, scheduling, and built‑in fault tolerance.
- Anyscale on Azure delivers production-grade cluster orchestration and job queue management to maximize GPU utilization across distributed AI workloads. Build your applications on a fully managed Ray lifecycle on AKS with priority-aware scheduling and Ray expert support ensure reliable, high-performance operation at scale.
- Run AI in your account with integrated Azure data, compute, governance, and billing. Anyscale on Azure is integrated with Azure security, policy, and identity patterns already deployed across the enterprise (including Entra/RBAC and private networking constructs) while running Ray workloads at scale on AKS.
Use cases
Explore Anyscale on Azure
Multimodal data preparation
Build and run scalable pipelines to curate and prepare multimodal datasets for foundation model training with Ray on Anyscale.
Distributed training & fine-tuning
Scale training from one to thousands of GPUs using your ML framework of choice with Ray on Anyscale.
Composite AI serving and inference
Deploy and orchestrate complex AI applications in production with Ray on Anyscale.
Embedding generation
Build efficient, scalable embedding pipelines using unified CPU preprocessing and GPU inference with Ray on Anyscale.
Reinforcement learning for LLMs
Scale RL post-training from a single node to thousands of GPUs with Ray, the engine for veRL, skyRL and more.
Physical AI simulation
Scale distributed simulation across thousands of concurrent environments to generate the training data and rollouts that autonomy and physical AI models need with Ray on Anyscale.
Industry scenarios
Learn how Anyscale on Azure powers your Python AI lifecycle
Perception & autonomous AI pipelines
Train and deploy multimodal AI across LIDAR, sensor, image, and video data to power autonomous vehicles, physical robots, and real-time control systems.
Real-time ranking and personalization
Run recommendation, ranking, and demand pipelines across text, image, and behavioral data to drive conversion and personalization at scale.
Low-latency risk modeling
Build and scale fraud detection, credit scoring, and prediction models on sequence and embedding data with speed and reliability.
Generation, search, and moderation
Power batch inference and training pipelines for search, generative AI features, and content moderation across text, image, audio, and video data.
Training, simulation & molecular inference
Train and fine-tune protein folding and molecular design models (AlphaFold, Boltz, RFDiffusion) across sequence, structure, and imaging data.
Read our announcement blog
Anyscale on Azure brings the power of Ray’s distributed execution together with Azure Kubernetes Service to create a purpose-built engine for modern, large-scale AI systems.
Empowering customers building the next generation of AI
Anyscale on Azure helps AI teams move from a single-machine experiment to production-grade systems running across hundreds or thousands of CPUs and GPUs.
-
Xoople builds Earth’s System of Record with planetary‑scale AI
"With Anyscale on Azure, Xoople can reliably run massively distributed AI workloads over planetary-scale satellite imagery, transforming complex spectral data into decision-ready intelligence.” - Milos Colic, VP of Engineering, Xoople -
Wayve rewrites the self-driving playbook
"Wayve and Microsoft have a deep partnership focused on scaling embodied AI and the infrastructure behind it. As Wayve’s AI platform and data operations have grown, Azure has become a core part of its large-scale compute and ML stack.” - Girish Venkataramani, VP of Engineering, Wayve AI -
Anyscale brings the power of Ray to Azure
"At Anyscale, we’re building a company around Ray. Ray is the open-source distributed compute engine for scaling python and AI workloads. What we’re bringing with Anyscale is the production layer on top: a performance-enhanced runtime, developer tooling, and managed operations so teams can stay focused on building. And now this is all delivered in an Azure-native manner with enterprise readiness on AKS.”- Robert Nishihara, Co-founder, Anyscale
Get the Azure mobile app 
