So far, in this book, we’ve been doing solo work, in which you and I are independently running code in a notebook to test different feature engineering algorithms and techniques to try and make the best pipeline we can. At the end of each case study, we’ve gotten to a place where we are generally happy with our results. Let’s say you are working on a project and are on to something. You want to see what it would take to get your ML pipeline and your feature engineering work into a production-ready state. You also want to bring in a trusted partner to help you continue the work, and you want to know how to enable them in any way you can, but all you have is a notebook with code that looks promising.
The next step in pushing your project forward is to consider modern data science and engineering practices to help you collaborate with new team members and to keep your data consistent and easy to use outside of your local development environments and notebooks. In this chapter we look at modern MLOps practices and, specifically, how to deploy and use a cloud-enabled feature store to store, handle, and distribute data.