Practical patterns for scaling machine learning from your laptop to a distributed cluster.

Distributing machine learning systems allow developers to handle extremely large datasets across multiple clusters, take advantage of automation tools, and benefit from hardware accelerations. This book reveals best practice techniques and insider tips for tackling the challenges of scaling machine learning systems.

In Distributed Machine Learning Patterns you will learn how to:

• Apply distributed systems patterns to build scalable and reliable machine learning projects
• Build ML pipelines with data ingestion, distributed training, model serving, and more
• Automate ML tasks with Kubernetes, TensorFlow, Kubeflow, and Argo Workflows
• Make trade-offs between different patterns and approaches
• Manage and monitor machine learning workloads at scale

Inside Distributed Machine Learning Patterns you’ll learn to apply established distributed systems patterns to machine learning projects—plus explore cutting-edge new patterns created specifically for machine learning. Firmly rooted in the real world, this book demonstrates how to apply patterns using examples based in TensorFlow, Kubernetes, Kubeflow, and Argo Workflows. Hands-on projects and clear, practical DevOps techniques let you easily launch, manage, and monitor cloud-native distributed machine learning pipelines.

About the technology

Deploying a machine learning application on a modern distributed system puts the spotlight on reliability, performance, security, and other operational concerns. In this in-depth guide, Yuan Tang, project lead of Argo and Kubeflow, shares patterns, examples, and hard-won insights on taking an ML model from a single device to a distributed cluster.

About the book

Distributed Machine Learning Patterns provides dozens of techniques for designing and deploying distributed machine learning systems. In it, you’ll learn patterns for distributed model training, managing unexpected failures, and dynamic model serving. You’ll appreciate the practical examples that accompany each pattern along with a full-scale project that implements distributed model training and inference with autoscaling on Kubernetes.

What's inside

• Data ingestion, distributed training, model serving, and more
• Automating Kubernetes and TensorFlow with Kubeflow and Argo Workflows
• Manage and monitor workloads at scale

About the reader

For data analysts and engineers familiar with the basics of machine learning, Bash, Python, and Docker.

Table of Contents

PART 1 BASIC CONCEPTS AND BACKGROUND

1 Introduction to distributed machine learning systems

PART 2 PATTERNS OF DISTRIBUTED MACHINE LEARNING SYSTEMS

2 Data ingestion patterns

3 Distributed training patterns

4 Model serving patterns

5 Workflow patterns

6 Operation patterns

PART 3 BUILDING A DISTRIBUTED MACHINE LEARNING WORKFLOW

7 Project overview and system architecture

8 Overview of relevant technologies

9 A complete implementation

About the Author

Yuan Tang is currently a founding engineer at Akuity. Previously he was a senior software engineer at Alibaba Group, building AI infrastructure and AutoML platforms on Kubernetes. Yuan is co-chair of Kubeflow, maintainer of Argo, TensorFlow, XGBoost, and Apache MXNet. He is the co-author of TensorFlow in Practice and author of the TensorFlow implementation of Dive into Deep Learning.