Both are good recommendations. Building ML systems isn't really about building ML systems, it's more about introductory machine learning techniques using Python

Both are good recommendations. Building ML systems isn't really about building ML systems, it's more about introductory machine learning techniques using Python libraries, and covers much of the same ground. I found Programming Collective Intelligence more practical, and the copy of Building Machine Learning Systems I have does have quite a few errors in it, which I suppose makes for more interesting experimentation (fixing example code certainly helps you understand it, no?).