We've had an interesting experience on the interviewing side of this. Asking a system design question and getting answers involving multiple layers of caching, pub/sub, event-driven whatever/etc.. when the real answer is to just use postgres. It handles the scale we're asking about. We know it, as that's what we use internally.

I've only worked at my startup so I can't comment on scale elsewhere, but if our simple architecture can handle 30k requests per second, I think it can handle most other companies scale too.