Uploaded December 2025 | Updated September 2026, 2 weeks ago
At Ray Summit 2025, we took a walk around the expo floor to get a quick, engineer-led overview of the Ray libraries built on Ray Core.
Ray is an open source AI Compute Engine for scaling AI and Python applications like machine learning. With Ray, developers can build and run distributed applications - without any prior distributed systems expertise.
đź”— Learn more about Ray: ray.io
In this video, we stop by the booths to hear directly from the engineers building Ray and walk through how the Ray ecosystem fits together — from data processing and distributed training to model serving and Kubernetes.
Whether you’re new to Ray or already using it in production, this walkthrough gives a quick look at what each Ray library is designed for and how teams use them in practice.
Ray Libraries featured in this video:
Ray Data - Scalable data processing for ML and AI workloads
Ray Train - Distributed training and fine-tuning
Ray Serve - Scalable model serving for online inference
RLlib - Scalable reinforcement learning
KubeRay - Running Ray on Kubernetes
Chapters:
00:00 Introduction and Welcome
00:17 Ray Data
01:48 Ray Train
02:57 Ray Serve
05:18 KubeRay
06:00 RlLib
At Ray Summit 2025, we took a walk around the expo floor to get a quick, engineer-led overview of the Ray libraries built on Ray Core.
Ray is an open source AI Compute Engine for scaling AI and Python applications like machine learning. With Ray, developers can build and run distributed applications - without any prior distributed systems expertise.
đź”— Learn more about Ray: ray.io
In this video, we stop by the booths to hear directly from the engineers building Ray and walk through how the Ray ecosystem fits together — from data processing and distributed training to model serving and Kubernetes.
Whether you’re new to Ray or already using it in production, this walkthrough gives a quick look at what each Ray library is designed for and how teams use them in practice.
Ray Libraries featured in this video:
Ray Data - Scalable data processing for ML and AI workloads
Ray Train - Distributed training and fine-tuning
Ray Serve - Scalable model serving for online inference
RLlib - Scalable reinforcement learning
KubeRay - Running Ray on Kubernetes
Chapters:
00:00 Introduction and Welcome
00:17 Ray Data
01:48 Ray Train
02:57 Ray Serve
05:18 KubeRay
06:00 RlLib
