Anyscale Courses

Optimizing LLM Inference with AWS Trainium, Ray, vLLM, and Anyscale

Explore cost-effective LLM inference scaling using AWS accelerators, Ray, vLLM, and Anyscale. Learn to build a complete stack on EKS and leverage advanced cluster management for enterprise-grade GenAI workloads.

• YouTube
• 46 minutes
• On-Demand
• Free Video

Scalable and Cost-Efficient AI Workloads with AWS and Anyscale

Learn to scale and productionize GenAI and LLM workloads cost-effectively using AWS compute instances and Anyscale's capabilities for ambitious AI projects.

• YouTube
• 35 minutes
• On-Demand
• Free Video

Fast and Scalable Model Training with PyTorch and Ray

Learn distributed model training with PyTorch and Ray. Migrate code, scale AI workflows, and optimize performance for large-scale training and fine-tuning on Anyscale.

• YouTube
• 58 minutes
• On-Demand
• Free Video

End-to-End LLM Workflows with Anyscale

Explore building and scaling end-to-end LLM workflows, covering data processing, model fine-tuning, evaluations, and production inference with Anyscale's modern platform.

• YouTube
• 46 minutes
• On-Demand
• Free Video

Building a Chat Assistant with Canopy and Anyscale Endpoints

Explore building RAG-based chat assistants using Canopy and Anyscale Endpoints. Learn architecture, see live examples, and get started with your own project using flexible frameworks and serverless LLM APIs.

• YouTube
• 45 minutes
• On-Demand
• Free Video

Ray Train: A Production-Ready Library for Distributed Deep Learning

Explore Ray Train's architecture for efficient, cost-effective distributed deep learning. Learn about resource scheduling, API simplicity, and exclusive features for LLM training, including Distributed Checkpointing.

• YouTube
• 32 minutes
• On-Demand
• Free Video

Redesigning Scheduling in Ray to Improve Cost-Efficiency at Scale

Explore Ray scheduling features to optimize AI applications, enhancing performance and cost-efficiency. Learn placement groups, graceful node draining, and label-based scheduling for faster, cheaper operations.

• YouTube
• 30 minutes
• On-Demand
• Free Video

Ray Scalability Deep Dive: The Journey to Support 4,000 Nodes

Explore Ray's scalability improvements post-2.0, including health checks and resource broadcasting. Learn to develop scalable code for large-scale ML workloads and understand challenges in building a 4000-node cluster.

• YouTube
• 31 minutes
• On-Demand
• Free Video

Scaling Time-Series Forecasting Models to Cope with the Multi-Verse with Ray

Learn to scale probabilistic time-series forecasting using Ray for financial markets. Explore techniques for handling non-stationarity and improving model robustness through back-testing and distributed computing.

• YouTube
• 25 minutes
• On-Demand
• Free Video

Developing and Serving RAG-Based LLM Applications in Production

Learn to develop, evaluate, and scale RAG-based LLM applications for production, including advanced topics like hybrid routing to bridge the gap between open-source and closed LLMs.

• YouTube
• 29 minutes
• On-Demand
• Free Video

AI Factory: Accelerating Solutions with Ray

Explore Lockheed Martin's AI Factory ecosystem for training, deploying, and sustaining AI solutions. Learn how Ray enhances ML workloads, enables distributed computing, and integrates with various tools to accelerate AI development.

• YouTube
• 28 minutes
• On-Demand
• Free Video

Deploying Many Models Efficiently with Ray Serve

Explore efficient deployment of multiple models using Ray Serve's features: model composition, multi-application, and model multiplexing. Learn industry patterns and case studies for optimizing resource utilization.

• YouTube
• 26 minutes
• On-Demand
• Free Video

From Spark to Ray: An Exabyte-Scale Production Migration Case Study

Explore Amazon's exabyte-scale migration from Spark to Ray, covering challenges, strategies, and future vision for integrating Ray into critical data pipelines.

• YouTube
• 33 minutes
• On-Demand
• Free Video

Modernizing DoorDash Model Serving Platform with Ray Serve

Explore DoorDash's journey in modernizing their model serving platform using Ray Serve, focusing on flexibility and self-service for diverse ML applications and frameworks.

• YouTube
• 31 minutes
• On-Demand
• Free Video

Scaling up Terascale Deep Learning on Commodity CPUs with ThirdAI and Ray

Explore ThirdAI's BOLT engine for efficient deep learning on CPUs, leveraging Ray Core for distributed training. Learn about overcoming communication bottlenecks and achieving near-linear scaling for terabyte-scale datasets.

• YouTube
• 26 minutes
• On-Demand
• Free Video