Reactive machine learning is an approach to building modern machine learning systems that I developed over the course of several years. The canonical treatment of the topic is the book.  Many of my talks, articles, and interviews touch on the idea. Since completing the book, I've focused on implementing some of the ideas of reactive machine learning in the Galápagos Nǎo project. And believe it or not, this whole body of work got started with this surprisingly popular blog post about dog poop.