Backbone workflow for ML Pipeline

[chadac]

Think this one merits some discussion. In the past, I've broken up my pipelines into stages where each stage usually has an entrypoint named "run" or something similar, accepting a small set of parameters to configure and test each process independently from a sample notebook, but still broken up into pieces so deep dives from notebooks are simple as well. But I've found that this can be a bit confusing as well on the execution end, and it ended up being fairly restrictive when I wanted to rethink the pipeline structure.

[Tsmith5151]

See response here. Might be some overlap where this could all be included in the starter notebook. As this evolves into other repos that provides a more deep dive into the ML pipeline, this would warrant its own notebook and thus we would just point back to the notebook w/ recommended best practices on getting started.