While working on the client projects there were several hiccups that would divert my focus from the actual insight generation.
For instance,
1. Limitation on the processing power
2. Missing track of trained ML model version
3. Managing the python packages and other open-source tool integrations

and more.
With a lot of overhead taken care by the Digital Hub, my time is well utilized. Check out this article on how DigitalHub can eliminate these roadblocks.

https://www.digitalhub.io/post/we-needed-better-and-it-didn-t-exit-so-we-built-it-an-open-source-data-science-ecosystem-part-1