A web-friend of mine just interviewed for a tech lead position at Google. Here’s a (slightly tidied up and anonymized) version of their experiences:
Most of my work, at least at the start, should be in “production software”–googlese for the software that helps keep Google’s amazingly huge distributed system running smoothly and seamlessly, and is mostly Python though with ample helpings of C++ here and there and a little bit of Java where integration is needed with some Java-centric application server (e.g. to serve google-ads on sites using such servers).
Plenty of “sideshows” doing such things as statistical analysis and data mining on the huge wealth of data Google collects, maybe giving [name deleted]‘s team a hand in data-quality assurance, etc, etc. Plus, every Google techie is supposed to use 20% of his time working on his or her own pet projects which might become Google’s Next Big Thing—that’s how gmail was born.
The selection process is grueling—multiple rounds of phone interviews where they ask you (depending on the fields of expertise you claim) everything from what’s 210, to how you would tweak bits in C to find out if a machine’s stack grows up or down in memory, all the way to having you “program on the phone”… then all of a sudden they rush you to Silicon Valley and you get a long full day of nonstop interviewing. I didn’t quite ace mine because I hadn’t thought of cramming on TCP/IP fundamentals, so I didn’t remember which bits are on in the three packets of the handshake (it’s
ACK—I could have worked it out, but not jetlagged and after about 6 hours’ interviews .
I made up for that when they had me program at the whiteboard a C++ implementation of unbounded precision multiplication; I did a test-driven implementation of the trivial routine with
std::vector<digit> containers, then did some handwaving about the Karatsuba algorithm (far too hard to implement standing up at a whiteboard, of course and could sense I had struck lucky… The guy interviewing me at that time had never really done unbounded precision computation work (at least not implementation of high-quality libraries for it), so by just opening the door a crack to the huge and mathematics that underlies that field (in which I had the good fortune to dabble a bit — a byproduct of my interests in combinatorial arithmetic) I had apparently exceeded expectations.
Lots of back-of-envelope computation and the like, too. A friend of mine thought he was doing well in his second Google phone interview when asked to sketch a way to compute bigram statistics for a corpus of a hundred million documents—he had started discussing
std::map<std::string> and the like, and didn’t get why the interviewer seemed distinctly unimpressed, until I pointed out even if documents are only a couple thousand words each, where are you going to STORE those two hundred billion words—in memory?! That’s a job for an enterprise-scale database engine!
So, at least as far as the interviewing process goes, it seems designed for people with a vast array of interests related to programming, computation, modeling, data processing, networking, and good problem-rough-sizing abilities—I guess Google routinely faces problems that may not be hugely complex but are made so by the sheer scale involved. I can just hope the actual day-to-day work is as interesting, fascinating and challenging as the interviews were—but from all I hear, it probably is. And they have bar-quality espresso machines in rest areas…