Pros
The company has given me opportunities to work on problems which have significant impact in many domains, some of which I wouldn't have imagined I'd work on when I was a university student.
Researchers are given a great deal of freedom in the projects - both in terms of how the challenges of an active project can be approached as well as what projects the company undertakes. Rather than having projects passed down 'from higher up', as many companies might do, everyone is involved in the scoping and planning of a project right from the start. This gives enormous freedom and opportunity to the team in identifying projects which are challenging, impactful and of interest to everyone.
As a result projects are not academic style work refining someone else's ideas or improving some software widget - they have scope and impact. Rather than trying to apply a single idea to multiple projects each project is mathematically and qualitatively unique. This means constantly developing new ideas and mathematical skills, instead of rehashing the same ideas every project. It means the work never gets 'easy' but the outcomes, both individually and as a company, are worth it.
Day-to-day work is a combination of individual work, team work and broader discussions. Many mathematicians have a preferred working style and the 'jolt' of becoming part of a team with regular updates and discussions can be difficult for a while the end result is always to improve someone's communication, idea generation and analytic thought processes. The company's approach to doing research is different and takes some time to adjust to - bad habits from a PhD die hard. But any interface with Industry - which my time at the company has given me many opportunities to do - will highlight the need for new approaches, not more of the same.
I doubt many companies have team-wide discussions which can range from subtle points of mathematics all the way to philosophy, let alone encourage them. Some could view such matters as 'distractions' or 'unnecessary' but I've found them rewarding over the years, even if I've been wrong in a fair few of them.
In addition to the flexibility of mathematical approaches for projects everyone in the company is given opportunities to meet clients or attend events, if they wish. This can range from be part of in-office events through to attending conferences elsewhere in the world.
Work at the company can be hard or challenging but anything worthwhile is and the work environment, challenges and opportunities make it worthwhile.
Cons
The biggest challenge is recruitment.
Many either do not want or are not up to the kind of challenges and problems we work on, which our interview procedure is the merest hint of. Others find adjustment from the most familiar 'academic style' too much.
I saw a similar thing at PhD - people realising 'research' wasn't as fun as 'lectures' and felt stopping was the right course of action, despite initial enthusiasm.