We're evaluating building our own ML file/data versioning/caching solution or using an open source solution. Anyone have experience with DVC? https://dvc.org/ they use Git for data version control, which is as exciting as it is terrifying. Any other suggestions? DeltaLake looks positioned for huge enterprise, not ML startups.

DVC have a nice community and team is very responsive. Git is a nice choice, however there is no visualisation.