Registration + Networking

Training Kickoff & Environment Setup

Learn Ray’s core concepts, including tasks, actors, and clusters. Understand how Ray scales Python and ML workloads, manage resources, and run distributed programs reliably across nodes.

Module 1: Scaling Python for AI Workloads

Explore how to build scalable data pipelines. Learn to ingest, transform, and preprocess large, multimodal datasets efficiently to feed training and inference workloads. Build streaming pipelines that chunk documents, generate embeddings at scale, and populate vector stores to power RAG workflows.

Module 2: Building Scalable Data & RAG Pipelines

Break

Learn how to scale model training using different model and data parallelism techniques like FSDP. Configure distributed training jobs, manage checkpoints, and integrate popular ML frameworks (like PyTorch) while maintaining reliability and performance at scale.


Module 3: Distributed Training at scale

Build and deploy scalable inference services. Learn request routing, autoscaling, and model composition patterns to serve models reliably in real-world production environments.

Module 4: Serving models in Production

 Q&A + Closing Remarks

 Happy Hour & Networking

Training kickoff and environment setup

Module 1: Scaling Python for AI Workloads with Ray

Overview of Ray's core concepts: How Ray provides simple, unified APIs for cluster computing. Ray Libraries. Setup and basic execution examples.

Module 1: Ray, the Foundation for Distributed Physical AI

Ingest, transform, and preprocess large multimodal datasets, then create streaming pipelines to chunk data, generate embeddings at scale.

Module 2: Building scalable multimodal data pipelines

Overview of data preprocessing and distributed training libraries, Ray Data & Ray Train, with an emphasis on vision language action models.

Module 2: Large scale VLA fine-tuning

Scale model training with data and model parallelism (e.g., FSDP). Run distributed jobs, manage checkpoints, and integrate frameworks like PyTorch with reliability and performance.

Using Ray Core to parallelize computationally intensive simulations (e.g., Mujoco, Isaac Sim)

Module 3: Robotics simulation

Q&A and closing remarks

<strong>Happy Hour + Networking</strong>

Agenda at a glance

Join us for Ray Workshop Boston on May 19, 2026 at convene, one boston place. Learn directly from Ray’s creators and power builders at this free technical hands-on workshops designed to help you move your AI workloads from experiment to production.

Ray Workshop is completely free to attend—our way of giving back to the Ray community.

Ray Workshop is a  fully in-person experience. We’d love to see you there!

If you’re interested in sponsoring Ray Day, please email [colette@anyscale.com](mailto:colette@anyscale.com).

Yes! Ray Workshop is also coming to Seattle, New York, London, Boston and San Francisco. Stay tuned for more details and upcoming announcements.

If your plans change, please let us know by emailing [events@anyscale.com](mailto:events@anyscale.com) so we can open your spot to someone else.

Anyscale logo

Ian Jordan

Alicia Chua

Suman Debnath

Pawarit Laosunthara

Featured Speakers

*Scale AI with confidence— Learn directly from Ray’s creators and power builders at this free technical hands-on workshops designed to help you move your AI workloads from experiment to production.*

boston-hero

What to expect

boston-place