It's a cross between grad school and a hundred little startups. I haven't always gotten guidance from management about what's important or how the teams need to work together. Like grad school, there's times where it does feel all your responsibility. Marketing and bigger vision sometimes comes from the product managers, but it always feels like individual advice rather than a single clear vision of where we should go.
Individual teams have a lot of control over libraries and code they use, so lots of infrastructure projects grow as research projects that succeed only if adopted by significant numbers of other teams.
Although there is a big vision for the company, it isn't as focused or controlled as in other companies; there's really an assumption that the right stuff will bubble up. It's not a place with the razor-focused direction.
Initial titles/ranks and promotions are determined by committees of other engineers. This is great because you're recognized for your engineering work, but bad if you aren't churning out enough code or if you're not having enough impact on the rest of the company. Initial titles get assigned 6-12 months in when you're put in the same pool as existing Googlers who are up for promotion. If you don't match up to them, you go down a slot - no difference in salary, bonuses based on new level, and any mental scars from being judged unworthy. It doesn't really matter, but if you're at Google you're probably not used to failing.
Everyone's driven to succeed. There may not be a lot of external pressure from management to pull long hours, but folks tend to do it anyway because they want to accomplish something great. It's an easy place to feel you're below average, even when you've been tops everywhere else.
