Anyscale Courses

Operationalizing Ray Serve on Kubernetes

Discover a new declarative REST API for Ray Serve, enabling seamless configuration and updates of applications within your MLOps lifecycle on Kubernetes.

• YouTube
• 30 minutes
• Self-Paced
• Free Video

Large-Scale Deep Learning to Augment Production RL Workloads at Riot Games

Learn to leverage large-scale deep learning for building game bots, predicting outcomes, and optimizing training processes in game development. Explore supervised learning techniques and scaling with Ray tools.

• YouTube
• 31 minutes
• Self-Paced
• Free Video

Cruise.data - A New Dataset Processing Pipeline for Cruise ML

Explore Cruise.data, a novel ML data pre-processing framework combining tf.data, PyTorch, and large-scale processing capabilities for efficient and scalable dataset handling in autonomous vehicle development.

• YouTube
• 31 minutes
• Self-Paced
• Free Video

Production RL and Decision-Making with RLlib

Learn to implement production-ready reinforcement learning and decision-making systems using RLlib, exploring real-world applications, challenges, and best practices.

• YouTube
• 29 minutes
• Self-Paced
• Free Video

Large-Scale Distributed Training with TorchX and Ray

Learn to launch elastic large-scale distributed training jobs using TorchX and Ray, overcoming traditional barriers and simplifying the transition from research to production.

• YouTube
• 31 minutes
• Self-Paced
• Free Video

Gearing Up for FIFA 2022 Using RLlib-Powered Traffic Control

Optimize traffic control for FIFA 2022 using RLlib-powered multi-agent reinforcement learning, microsimulation, and Graph Convolutional Networks for congestion management in Qatar.

• YouTube
• 33 minutes
• Self-Paced
• Free Video

State of Ray Serve in 2.0 - Features and Updates for Multi-model Inference

Explore Ray Serve 2.0's features, use cases, and architecture for multi-model inference and composition. Learn about autoscaling and production hardening techniques.

• YouTube
• 32 minutes
• Self-Paced
• Free Video

Highly Available Architectures for Online Serving in Ray

Explore highly available serving in Ray 2.0, its architecture, functionality, and practical deployment for improved efficiency and reduced disruption during failures.

• YouTube
• 34 minutes
• Self-Paced
• Free Video

FastCFE - A Counterfactual Explainer Using Reinforcement Learning

Explore reinforcement learning for real-time counterfactual explanations in machine learning explainability. Learn about FastCFE algorithm, OpenAI Gym, and Ray+RLlib for actionable insights in various applications.

• YouTube
• 31 minutes
• Self-Paced
• Free Video

A Serverless Resilient Graph Analysis Platform Built on Ray - ByteGAP

Explore ByteGAP: a serverless graph analysis platform for processing super large-scale graphs, built on Ray at ByteDance. Learn about its architecture, features, and advantages.

• YouTube
• 34 minutes
• Self-Paced
• Free Video

Scalable Feature Engineering with Hamilton on Ray

Explore scalable feature engineering using Hamilton and Ray at StitchFix. Learn declarative dataflow, team scaling, and efficient computation for thousands of time-series features.

• YouTube
• 30 minutes
• Self-Paced
• Free Video

What's New in RLlib 2.0 - Key Improvements and Features

Explore RLlib 2.0's improvements: configuration, algorithms, customization, policy serving, and Ray AIR integration for intuitive and performant reinforcement learning.

• YouTube
• 28 minutes
• Self-Paced
• Free Video

Distributed Training with Ray on Kubernetes - Lyft's ML Platform

Leverage Ray on Kubernetes for distributed training at scale. Learn to create on-demand clusters and simplify ML workflows with custom SDKs.

• YouTube
• 32 minutes
• Self-Paced
• Free Video

Zero-Copy Model Loading with Ray and PyTorch for Efficient Deep Learning Inference

Optimize deep learning model loading for production using PyTorch and Ray. Implement zero-copy techniques to reduce costs and improve performance in NLP model deployments.

• YouTube
• 33 minutes
• Self-Paced
• Free Video

Reducing Cost, Latency, and Manual Efforts in Hyperparameter Tuning at Redicell

Optimize machine learning model training with Ray Tune. Reduce costs, latency, and manual efforts through efficient hyperparameter tuning, resource management, and experiment tracking.

• YouTube
• 31 minutes
• Self-Paced
• Free Video