Pros
I thought this was a cool place to work. Had some run-of-the-mill business-oriented problems, but also some unique opportunities to innovate and work on some interesting, creative projects.
Cons
Though there was a definite appreciation and excitement about building out the data science team, there was also a rapid decline in data science understanding among the higher-ups. This would sometimes result in oddly constrained requests where the stakeholder wanted the team to use machine learning and pattern recognition techniques to identify natural clusters of the customer base (as opposed to the rules-based definitions already in use), but that there must be exactly N clusters, all more-or-less of size X, and that we must use variables a, b, and c. Constraints are fine, but it sometimes felt like "what's the point?" if we're not going to optimize the hyperparameters, or let the model dictate which variables are most important and should be used, etc. This wasn't a huge issue, but it did require good communication and negotiation skills in order to help them get more value out of the project.