As a Data Analyst, I'd like to more easily branch the code and flow of my project, make changes on the branched version, and then re-integrate those changes back into the original project. This would more easily allow project development while keeping the current project(s) in production use. This makes it more likely that more participants in the project can apply project development best practices.
Dataiku is more explicit about best practices for merging branched projects back into an existing project. Currently, most documentation says that this is up to the practitioner to figure out.
This merging process can be done all inside Dataiku DSS
There are several threads here on the Community about these challenges.