Sounds like a fraternity hazing ritual to me. I'm curious your target have access to Google.com during your "interview"? Also what is going to happen to your hazing process when people start walking in with Google glass? If your canned intelligence test questions gave been recorded and indexed aren't you going to feel a little silly?
If a company is using f2f interviews to "weed" out candidates they either have a very broken recruitment team or really think so highly of themselves that they are in need of a wakeup call.
For having such a reputation as an innovative company it sure seems odd to rely on an antiquated process that seems straight out of the social eugenics movement of the 1930s.
> I'm curious your target have access to Google.com during your "interview"?
No, of course they don't. The idea is to see how well they think without being able to "just google it", so when they're tackling a problem nobody has tackled before, they aren't completely lost.
> Also what is going to happen to your hazing process when people start walking in with Google glass?
Probably they'll get asked to take them off, much the same as you might do to a candidate who walked in using a cellphone.
> If your canned intelligence test questions gave been recorded and indexed aren't you going to feel a little silly?
Candidates are told that the interviews are confidential. And no, I wouldn't feel silly if people recorded the questions and put them on an internet. It would be kind of a warning sign if we were embarrassed about them - the reason they're confidential at the moment is the same reason you can't look up upcoming exam papers on the internet.
> If a company is using f2f interviews to "weed" out candidates...
What on earth is a job interview for if not to reject some candidates? It's kind of a negative way of looking at it, but that is exactly what they're there for; to weed out candidates who are good (they must be at least adequate to get that far) but not good enough at the moment.
I'm not even gonna start on the last paragraph. Goodness knows nobody, especially Google, have a perfect interview process, but comparing it to eugenics is a little hysterical.
Umm, a common practice in the US in the 1930s at state fairs was to take IQ tests in order to become informed about your "feebleness" so as to aid you in your decision and fitness for marriage and procreation. In some jurisdictions, you could not enter into marriage if you were "feeble-minded." [0]
Most Americans don't realize there were several programs of enforced, mandatory sterilizations of the "feeble" as part of the social eugenics craze in the US. Faribault, Minnesota was a pretty active center of such activity.
Anyway, I think the comparison is apt and illustrative. Using "intelligence" measures to quantify someone for something entirely irrelevant to the task being tested for. In the first case "thinking on your feet" questions and professional software engineering and in the second procreating and raising children free from genetic defect. Hopefully with a little reflection you will see the connection too.
Whew. Good to know, I'm glad you cleared that up. I'm assuming they also are not racist because that is part of the analogy carried to illogical extremes too, right?
You mention that the interview process is confidential and thus immune from recording/playback. Is that true? Do candidates sign something? In any case, I think your allusion to academia is interesting ("same reason you can't look up upcoming exams"). Isn't this entirely the wrong model for forecasting future job performance and the main complaint of most people about the undergraduate university system ("learn" in order to pass the test)? The original comment even mentions failure to "cram" prior to the interview as an indicator of insufficent enthusiasm.
Anyway, I think if a candidate came into an interview and needed google or stackoverflow or whatever to function, I would provide that as a resource and would use it as an opportunity so as to judge the dependence and quality of their workflow because it is not extraordinary to see how everyone (even brilliant geniuses I know!) use these resources on a daily basis. I don't think I would deny someone access to prescription medications affecting cognitive function/enhancement either nor could I even do so legally. How on the one hand can you use internet contributions (well-reputed blog posts, open source contributions, active social media following, etc) as positive evidence of candidate desirability and also at the same time view using same as negative?
I guess I would could care if I was for example screening someone that as part of their duties they were expected to say represent me speaking at a conference or that the work product was extremely confidential (something which would necessarily require curtailing access on the job).
Finally I have to imagine in the very near future if the current trend of viewing "internet access" as an universal human right continues and it becomes even more of a basic enabling technology of the human experience, denying someone access during a job interview is going to become a dicey proposition the same as not providing accomodations for and not taking into account the cost of say wheelchair access, etc is today. The irony of Google denying access to the internet during a job interview is not without some comedic value.
How is this a hazing process? The OP mentioned a number of ways he tries to simplify it or keep people from being intimidated by the technological implications. They are looking for people who can think on their feet.
To be honest your entire post sounds jealous / bitter.
A hazing process is a tribal ritual used to stress candidates and humiliate them to weed out those who are deemed unfit to some standard that members of the actual tribe usually don't conform anyway and to measure the devotion of the candidate to the tribe by submitting to and acquiescing to the authority of the ritual.
I have no idea why you would suggest jealousy. I have no intention of ever working for Google unless I am acquired. Perhaps my reply was poorly worded, but I find this kind of process very unenlightened. If I'm "bitter" it's only because I guess I would expect better from a company that has vacuumed up so much available oxygen. Also, I have no idea whether this is actually the case at Google, I was just responding to someone who represented their experiences as such.
Finally, the last thing you want from an engineer is to be "good on their feet" unless you are hiring for some critical ops position. Rather you want someone that takes measured analytical approach.
This is like choosing a President based on their debate performance or the security provisions of the TSA. It's just theater bordering on the absurd.
> I'm curious your target have access to Google.com during your "interview"?
I think it would have been a good idea for Google to publish online an official "Google interview cheat sheet" and hand a copy to the interviewee at the start of the interview.
However, if the candidates know beforehand that the web will be available, they're going to study less; it's just human nature. If they have 20 minutes for a question, and they have to use the web to recall the basis for the question, they're probably going to waste 5 minutes compared to someone who has taken the time to study.
> If your canned intelligence test questions gave been recorded and indexed aren't you going to feel a little silly?
They weren't intelligence questions. Many of brilliant people would have failed miserably and many of people with IQs lower than the Google median would do very well. Many of the questions I asked were simplified versions of problems I saw and fixed in Google's codebase. I wanted to make sure that people who got hired were capable of diagnosing and fixing the real kinds of problems I had seen.
Edit: spelling mistake, added "it's just human nature".
Note that I wrote the OP in response to someone who seemed to think they were better than having to memorize/review the basic data structures. The measures I took to reduce stress weren't relevant to why it was necessary to know how basic data structures work, and I didn't mention them.
As someone else has mentioned, the interview process may have changed significantly since I left in 2010. Clearly I don't think Google is perfect, but they try hard to make the interview process as accurate as possible. It's very expensive in terms of opportunity cost alone to interview someone, and every person they stress out in the interview process is going to tell at least 10 good potential hires about their bad experience.
Generally the first thing I did was tell the interviewee that every interview was a clean slate and that I didn't know how they had done in previous interviews and wouldn't tell later interviewers about their performance. I told them the hiring committee takes the independent reports from the various interviewers and throws out any outliers, so they should try and not worry about any one interview. I mentioned that the sheet of paper I had been passed listed which questions they had been asked, but not how the interviewer thought they had done. I explained that I would gradually modify the questions to make them harder and harder until we ran out of the time I had allotted for that particular question, and that I didn't expect them to be able to answer the most difficult parts of the questions in the allotted time. I mentioned that I was more interested in seeing how they thought than seeing one best answer.
I generally tried to structure the interview as a joint problem solving session, trying to use the word "we" as much as possible. I gave hints when people got stuck, and certainly didn't expect them to fully solve the problems without hints. In my writeup, I'd often note if I felt they would have done a bit better if they had been a bit calmer. I really was silently rooting for them to succeed. It's very fun interviewing someone who's bright and excited about the subject matter.
That being said, I had 40 minutes (trying to leave 5 minutes at the end for any questions they had) to try and asses the interviewee's abilities in several categories. I'm not sure how to do that without ratcheting up the difficulty of a question until the interviewee fails, and then repeating with another question. If you don't push them to the point of failure, you don't know where the edges of their abilities are. I tried to minimize the pressure, but in the end, I had to push them until they failed, and failed multiple times. Failure is naturally stressful, and there was only so much I could do to reduce that stress.
The interview process tried to minimize any biases of any one given interviewer, making each judgement as independent as possible, so that meant there needed to be many interviews. There's a lot of opportunity cost spent in interviewing people, and people get worn out if you keep bringing them back in, so that necessitated making each interview short.
I knew a fair number of full-time Googlers who started out as contractors. I'm sure already working at Google helped take a lot of the pressure off, but they still went through technical interviews.
Anyway, the interview process was far from perfect, I'm just saying that it's perfectly reasonable for professionals to be able to cram freshman/sophomore level data structures in order to interview.
Yes, it sounds like you took your role very seriously and the applicants were probably extremely grateful.
I just think the whole process as you recount it seems inefficent and probably as useful as a fraternity hazing at getting the desired result. That the process is daunting is used to screen out the undeseriables even when completing the gauntlet has very low truth value on the question of performance in the role. But I'm sure that Google has some A/B testing right? Like measure some candidates effectiveness on the job that didn't go through the process (like I dunno low badge numbers) against those that did go through the process. (I had to sneak that in, lol?)
Not a bad thing in an interview, measuring how the candidate responds to failure/blocking as long as it is done honestly because creative solutions and overcoming failure/roadblocks is pretty much the same thing in engineering.
I have to agree to disagree with you on your conclusion. Cramming data structures seems to me like a waste of time.
If I had the resources of a Google I would innovate a solution so that the interviewer can focus on the problem that she is actually trying to solve: determining fitness.
Compare this for example to the NFL combine. Or the amount of time and money put into farm systems, feeder teams, and scouting reports. Why not pay the candidates to solve an engineering problem and judge the output? I don't have the answer, but I'm fairly certain that there has to be a better way than stress traps and good-ole-boy networks.
I work very hard to tailor my interview questions to the actual background of each particular applicant so that they can in efffect ask their own questions and demonstrate ability (ala "oh that's interesting how did you solve that problem, and what if you did this instead how would that work in the design, here's a whiteboard show me, etc). It is very hard and I regret very much the "good hires" I have probably let go because they couldn't engage well with me or talk about their own experience well. This obviously takes time and effort so I depend on recruitment/referrals properly to weed out wastes of time.
Is anybody working on this problem? It seems clear google isn't and HR innovation is extremely profitable.
What makes you think google hasn't tried your solution, among others?
Google spends a large amount of time, energy, and money trying to figure out the best ways to hire candidates that are "fit", and tries numerous things (simultaneously, in fact) past "asking stress problems". The fact that most people go through the current interview process (which, btw, has visibly changed in the 6.5 years i've been there) does not mean they don't experiment, or use alternative methods. It's simply the most common.
The people adjusting the hiring process are those that passed it. It is a filtering system that picks out a particular type of person and reinforces it. I find it hard to believe that it will change in any way that adjusts the filtering system. People at Google might experiment, but it will be within the orthodoxies, e.g., choosing between hybrid or gasoline vehicles, rather than considering mass transit.
Err, no they aren't, because they aren't all (or even mostly, AFAIK) engineers.
It's certainly not the case that engineering and other positions have the same interview process.
I would also qualify that as "Google spends a large amount of time, energy, and money trying to figure out the best ways to hire candidates AT SCALE"
There are literally millions of resumes being submitted every year, and the cost of a false positive is high (saying "yes" to a bad hire is generally much worse than saying NO to a fantastic hire).
You are correct, I don't know and was perhaps overgeneralizing based on the belief that the parent was accurately representing his/her experience as common.
What can you tell me about how the process in your experience has changed in the last six years? Experiments, even social ones, have little value unless their results are shared! ;-)
I'm always looking to improve my HR skills. Hiring is so hard. It's my opinion that one of the reasons it is so hard is (primarily?) because of all the subterfuge/deception involved by both parties of the transaction. I wonder whether the experience of successful dating services have any insights to offer? Seems like a similar problem but without the messiness of physical attraction to get in the way.
If a company is using f2f interviews to "weed" out candidates they either have a very broken recruitment team or really think so highly of themselves that they are in need of a wakeup call.
For having such a reputation as an innovative company it sure seems odd to rely on an antiquated process that seems straight out of the social eugenics movement of the 1930s.