Pros
I get to work on the most interesting and challenging problems in data science. We do real predictive modeling, no "click-counting." I work with intelligent and thoughtful coworkers and managers. Managers encourage ownership of work/concepts and creativity; not much micromanaging or checking boxes happening on my team. There is room for growth and promotion (both technical and people-lead paths), and managers actively help you grow. Compensation can be great, but meaningful raises often require negotiation (which has been a surprisingly positive experience so far). Work-life balance here cannot be beat. The culture within my organization and especially my team cannot be beat.
Cons
There is some disconnect across organizations that slows the progress of research making it in to production. Several strategic shifts and re-orgs over the past few years have left us with some tech debt. The acquisition by Bayer has caused some headaches (not least of which is the IT integration), all of which has led to increased turnover. There is the usual political BS at higher levels (but ICs are mostly protected from this). HR is a pretty inaccessible org in my experience, and seems to be overworked and/or understaffed lately.