What this book covers
Chapter 1, Introducing MLflow, will be an overview of the different features of MLflow, guiding you in installing and exploring the core features of the platform. After reading this chapter, you will be able to install and operate your MLflow environment locally.
Chapter 2, Your Machine Learning Project, introduces the focus of the book. The approach of this book is to work through a practical business case, namely stock market prediction, and, through this use case, explore all the different features of MLflow. A problem-framing framework will be used to get you deeply familiar with the example used in the book. A sample pipeline will be created for use in the remainder of the book.
Chapter 3, Your Data Science Workbench, helps you understand how to use MLflow to create your local environment so that you can develop your machine learning projects locally using all the different features provided by MLflow.
Chapter 4, Experiment Management in MLflow, is where you will gain practical experience of stock prediction by creating different models and comparing the metrics of different runs in MLflow. You will be guided as to how to deploy a tracking server so that many machine learning practitioners can share metrics and improve a model.
Chapter 5, Managing Models with MLflow, looks at the different features for model creation in MLflow. Built-in models, such as PyTorch and TensorFlow models, will be covered alongside custom models not available in MLflow. A model life cycle will be introduced alongside the Model Registry feature of MLflow.
Chapter 6, Introducing ML Systems Architecture, talks about the need to architect machine learning systems properly and how MLflow fits in the picture of an end-to-end machine learning system.
Chapter 7, Data and Feature Management, introduces data and feature management. The importance of feature generation will be made clear, as will how to use feature streams to log model results with MLflow.
Chapter 8, Training Models with MLflow, is where the complete training pipeline infrastructure will be described and developed for the problem at hand, with the use of MLflow-specific features.
Chapter 9, Deployment and Inference with MLflow, is where an end-to-end deployment infrastructure for our machine learning system, including the inference component, will be exposed using the API and batch features of MLflow. The cloud-enabled features of MLflow will also be described.
Chapter 10, Scaling Up Your Machine Learning Workflow, covers integration with high-performance/big data libraries that allow MLflow systems to scale for large volumes of data.
Chapter 11, Performance Monitoring, explores the important area of machine learning operations and how to ensure a smooth ride for the production systems developed in the book using best practices and operational patterns.
Chapter 12, Advanced Topics with MLFlow, presents advanced case studies with complete MLflow pipelines. The case studies use different types of models from the ones looked at in the rest of the book to ensure a breadth of feature coverage for MLflow.