At my company we have normalized data warehouses, and data scientists query them to make big, static, denormalized tables to run through their analyses and models.

I thought this was a pretty standard, table stakes move in the analytics world.