Thoughts and tips on organizing models for a machine learning project

Hi sure, will do that, thanks!

update: done, here: Ideas, tips & packages: model tracking, persistence, model db/store with pipeline params/metadata, reproducibility?