If you want to setup Azure DevOps the way that works best with Git, what would it be?

Usually, "One repository per independently deployable product" is stated. What is one product? As Data Platform people, many of us are used to thinking of our (former?) Data Warehouse as a black box. But that's not what meant here!

We'll take inspiration from our software development peers and look how to break up our data platform projects in multiple smaller repositories and accompanying pipelines.

We'll look into the practice of breaking up your data project into multiple small repositories and accompanying pipelines. How does that work with schema changes? And how do we orchestrate everything in a neat way? Are there other best practices to keep in mind?