As an actuary, I can not recommend it to anyone with a technical background. I have degrees in mathematics and actuarial science, and I consider my actuarial science degree completely worthless. A mathematics and/or computer science degree will teach you more in terms of how to think than anyone with an actuarial science degree. If you’d like to take a major where your professors literally teach to the exam/test, the…
Technical ownership is distributed from the top down and can not be overridden from bottom-up process improvements. Even if it made sense to integrate two departments that regularly pass each other data, that will be fought because that interferes with someone's corporate fiefdom.
There is no clear high-level diagram of how the whole company functions. So unlike a web-app with a client-frontend-routing-backend-database pipeline that a smart 15 year old could learn to conceptualize, instead you're expected to learn the inter-process requirements in a piecemeal fashion, one detail at a time as it arises in a weekly task. In the best case, this big picture starts to emerge ~5 years into a full time job. In the more likely case, your tasks never expose you to some part of the company's operations and they remain a blackbox and you'll be unable to reason about.
While inter-process communication remains opaque, complexity within each process proliferates. If you do life insurance, you need to support every product you sell for fifty years, so you deal with often poorly documented product rules and hard core legacy systems. Note - you can't deprecate anything because must honor those complicated agreements made decades ago.
Basically the frustrations boil down to doing interesting systems/data work but with being forced to do it in a worst-practices fashion.