Machine Learning Model Serving Patterns and Best Practices

Deploy machine learning models to production with ease. Deploy models in the cloud and make it available in large organizations.

Key Features

• Learn the best practices about bringing your models to production
• Learn to use the available tools to serve ML models and what are the comparisons and contrasts between them
• Understand state of the art monitoring approaches for model serving implementations.

Book Description

It's clear that data scientists are in high demand. And with good cause. But the talent shortage is just one of the problems. One big challenge for data science teams is that very few models make it out of the lab and into production. Serving model enables data science or ML teams to bring their models to production. Most of the ML models are not deployed. So, ML engineers need to know the critical steps in serving the ML model.

Developers working with ML will be able to put their knowledge to work with this practical guide to model serving. The book provides a hands-on approach to implementation and associated methodologies that will have you up-and-running, and productive in no time. The book will not only be a repository of all the steps in serving the ML model but also a strong reference for developers, data scientists and AI/ML teams to have fun.

Complete with step-by-step explanations, practical examples and self-assessment questions, you will begin by exploring different model serving technique, and then serve your first model using these serving techniques

You will learn different model serving techniques, how to get up and running with these serving techniques.

By the end of the book you will be able to save and serve your model using state-of-the-art techniques.

What you will learn

• Model Serving, Patterns of Model Serving
• Understand serving Machine Learning Models using different techniques
• Learn different approaches to Stateless Serving.
• Implement advanced techniques of Batch model serving
• Learn the fundamental concepts in continued model evaluation
• Serving Machine Learning Model using Fully Managed Cloud Solution

Who This Book Is For

The audience of this book are Machine Learning Engineers/Scientists, Data Engineers/Scientists who want to bring their models into production. They are familiar with machine learning and have experience of using machine learning techniques but are looking for options and strategies to bring their models to production.

Table of Contents

1. Introducing Model Serving
2. Introducing Model Serving Patterns
3. Stateless Model Serving
4. Continuous Model Evaluation
5. Keyed Prediction
6. Batch Model Serving Pattern
7. Online Model Serving Pattern
8. Two-phase model serving pattern
9. Pipeline pattern
10. Ensemble Pattern
11. Business logic pattern
12. Tensorflow Serving
13. Rayserve
14. BentoML
15. Serving ML Model using Fully Managed Cloud Solution