The Ethos of Model Development Is Capturing Everything That Goes Into Your Model Definition

From a product perspective, kind of the ethos is let's capture everything associated with your model development life cycle. And we leverage tools like Docker and so on to help make this a reality. But one of the fundamental tenets there is that first we want to enable reproducibility across the various training cycles of these models. If I run the same model on the same data today as I did yesterday, I should give the same model parameters out.