Chapter 4: Data Management for Machine Learning
As a practitioner of machine learning (ML) solutions architecture, I often get asked to help provide architecture advice on data management platforms for ML workloads. While data management platform architecture is mainly considered a separate technical discipline, it is an integral part of ML workloads. To design a comprehensive ML platform, ML solutions architects need to be familiar with the key data architecture considerations for machine learning and know the technical design of a data management platform to meet the needs of data scientists and the automated ML pipelines. In this chapter, we will look at where data management intercepts with ML. We will talk about key considerations for designing a data management platform for ML. We will then deep dive into the core architecture components for a data management platform and the relevant AWS technologies and services for building a data management platform on AWS. Finally, you will...