These would be stupid questions even if they weren&#x27;t being administered by a non-programming recruiter.

While I think these things are super interesting trivia, I really want to ask the question: &quot;Why does a Director of Engineering for Google need to know this?&quot; That is not only a rather unusual question (within the scope of daily tasks for a programmer, these tend to be an important minority for a project), but it&#x27;s one that&#x27;s best settled empirically and within the context of your execution environment.There are a lot of different answers to the question depending on where you&#x27;re doing it and how you want to impact the machine.It&#x27;s a bizarre question to pose to Director level, because the proper response even for a technical director should be, &quot;The answer to that changes quickly, we should measure and check what our environment&#x27;s latest capabilities are and if they are reliable.&quot;I mean I&#x27;m all gung ho to program stuff, but I think that&#x27;d be a massive misapplication of my time with that title in my current job, and given my other responsibilities I&#x27;d only do it more slowly than someone with dedicated focus.

Hahaha, that is funny&#x2F;tragic. Outsourcing hiring a director of engineering to a moron with a cheat sheet! Your answers were great, though, and indeed on the systems I have tried the popcount() computation with the bit shifts you suggested is faster than using an 8 bit lookup table. (however the SSE4 popcnt (<a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;SSE4#POPCNT_and_LZCNT" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;SSE4#POPCNT_and_LZCNT</a>) is faster still).

Maybe they set unrealistic expectations to see if the applicant lies about meeting them and give wrong answers to see if the person acts like a smartass.

terrible interview questions, that&#x27;s why only technical people should be giving technical interviews. Most answers to questions are more nuanced than a straight quantifiable answer they have on a sheet of paper.

So much this. They would have missed Einstein &amp; co.

This right here is why Google hit the wall and stopped innovating. Which is sad because they seem to be trying really hard.

only some hash functions are fingerprints.  Other hash functions are the exact opposite of a fingerprint.

recruiters need to have a few levels of keywords they scan for.level 1: foo, bar, baz 
level 2: frobnitz, barfoo
level 3: 42, etcIn someone is using words from level 2 that work together in the ways laid out, they&#x27;re probably beyond level 1, and wouldn&#x27;t use the word &#x27;fingerprint&#x27; (in this case) - they&#x27;re giving more detail (and probably better) than what was being listened for.

I&#x27;ve found the best way to answer those is &quot;buzzword diarrhea&quot;first thing I do is try to get them to ask it a different way. then I play spew out every related buzzword possible as if its a question. Are you talking about .... oh, you mean .... I can usually get it in a few tries.

I&#x27;d have failed too, because I use a very particular way to describe a hashing function.A hashing function is actually a sorting function. It&#x27;s supposed to take an input space and sort it in an unpredictable and evenly distributed way across the output space. What&#x27;s more, neighboring points in the input space, no matter the sort used to determine proximity should not result in neighboring points in the output space.Fingerprinting is just an emergent value that comes from choosing fixed length hashes, and the fact that the mapping from input to output is stable.

On my phone screen a Google recruiter asked me &quot;how much is 2^24&quot;, and I knew the answer by heart and answered immediately.So he asked &quot;how did you figure this out so fast?&quot;. I told him I didn&#x27;t, I just remember all the &quot;important&quot; powers of 2. He said &quot;well... that&#x27;s not what I was looking for, I wanted you to calculate it, but... I guess a candidate who memorizes powers of 2 is a positive sign?&quot;. I passed.

I got an interview with Amazon, but failed the first round, because I did not do well on the reasoning part of the online test. You couldn&#x27;t skip questions, so I spent too much time on some of them and had to rush the ones at the end. They failed me despite getting the coding part 100% right.Oh well.

The inode question gave me flashbacks to my interview with Amazon. They wanted me to explain what a hash function is. I kept giving answers for about 3 minutes explaining hashing, common algorithms, reasons to use it and places it applies.Recruiter: &quot;I was looking for you to say it&#x27;s a fingerprint&quot;So I guess I was wrong, because despite explaining them in decent detail, I didn&#x27;t use the one keyword.

&gt; Given the complexity of the work I was doing at the time, the entire thing seemed ridiculous.I had the same a couple of years ago, at the time I was a Rails developer at a small startup. The first screening they asked me some basic technical questions (I guess it was a recruiter), then the second screening they asked me to walk through some data structures and algorithms off the top of my head.I knew they would do this, and hadn&#x27;t looked at any CS algorithms since university, so got a copy of Programming Pearls and studied through that every evening for the week before. I picked a few that I hoped would come up, and it turned out that&#x27;s what they asked for. I think first a linked list (maybe doubly linked?) and then a tree sort.I surprisingly passed, but the attitude of the interviewer really put me off. He said he knew Python, Java and Go - none of which I had used - and wasn&#x27;t too happy when I said I wanted to use Ruby (which the first guy I spoke to, said was fine). Then throughout the interview it seemed very much like he was fighting against me and trying to prove me wrong.After that I couldn&#x27;t be bothered any more, I didn&#x27;t really want a job at Google, it&#x27;s just a recruiter contacted me and I decided to try it out. I guess this style of interviewing must work for Google, but it&#x27;s just not the way I like a company to introduce itself to me. It just seems like they are approaching it with so much ego, as if I would be privileged to work there, but to me a job should be mutually beneficial.

I had almost exactly this same engineering test when google interviewed me in 2006. It was terrible, and left a bad taste in my mouth. Given the complexity of the work I was doing at the time, the entire thing seemed ridiculous.

I&#x27;ve had somewhat similar experience with other big companies. People wonder why I hate to change jobs.

Believe it. The depths of corporate absurdity know few bounds.

This is a phone screen - the test they make you answer before they take the time to schedule actual interviews.

Director of Engineering is a soft-skills position that involves communicating with a lot of non-technical people. It doesn&#x27;t surprise me at all that the initial screen would be a conversation with a non-technical person.

Is this for real? Wow.
Reads like someone spoofing the real Google interview.For &quot;Director of Engineering&quot; they have a technical illiterate (nothing wrong with that, in other contexts) check for literal matches of answers? 
the mind boggles.(I&#x27;m really having a hard time believing this is for real)

Sometimes I wonder if this approach has hurt Google&#x27;s ability to innovate.

This hiring test makes me cringe. How is this any better than taking a multiple choice test in college? These questions asked should invite back-and-forth conversation like the author is providing, especially when the company is seeking out highly qualified candidates.

I didn&#x27;t hear about the light bulb drop question until this thread, and was surprised how in-depth it can go.Can you really &quot;fail&quot; an interview on a question like that? I assume the person conducting the interview is reviewing your process of coming to an answer rather than getting the most correct answer - right?By most correct I found this article that broke down the question&#x2F;answer: <a href="https:&#x2F;&#x2F;pointlessprogramming.wordpress.com&#x2F;2011&#x2F;03&#x2F;11&#x2F;2-light-bulbs-and-a-100-story-building-solution&#x2F;" rel="nofollow">https:&#x2F;&#x2F;pointlessprogramming.wordpress.com&#x2F;2011&#x2F;03&#x2F;11&#x2F;2-ligh...</a>

It makes sense if they want a firewall from the public, a legally defensible position, while behind the scenes they hire via other channels ( close friends )

This is really weird and should not be happening. I am a software developer and do a lot of interviews including phone interviews.These kind of questions are not expected, but I would say the lightbulb one is not super horrible as long as it is the discussion that was valued and not the answer. I have gotten weirder questions for sure when interviewing.

&gt; They asked me the same questions back in 2012. I was applying for a Java developer position.I&#x27;ve has similar issues with other companies. Network developer? Great! Let&#x27;s ask some assembly &#x2F; Java questions!Uh... if you people are that stupid, I don&#x27;t want to work for you. Thanks. &lt;click&gt;

That&#x27;s really interesting that he said he was a psychologist.  I am probably giving them too much credit, but maybe they rare screening for personality with this test and not for technical knowledge.  If you handle a frustrating and difficult situation gracefully, perhaps how many answers you got &quot;correct&quot; is irrelevant?

<pre><code> &gt; They asked me the same questions back in 2012. I was applying for a Java developer position.
</code></pre>
OMFG.

They asked me the same questions back in 2012. I was applying for a Java developer position.The person asking the questions said right from the start that he is not technical (he was a psychologist) and that he had a cheatsheet in front of him.I passed that telephone call and failed the process later on another call (on the lightbulbs and the 100 floor building)Really surprised to see that 4 years later they still do this.

I bet you spent more than 30 minutes (typical phone screen length) looking at TLPs.  Spending 2 minutes to google DLLP&#x2F;TLP packet structure is not unreasonable.

It tends to be incredibly helpful to know details like the length of a mac address and what tcp flags mean what if you&#x27;re looking at an unformatted memory dump.One example of how this can be handy is when doing low-level NIC bring up, I&#x27;ve read PCIe TLPs from a logic analyzer to correlate them to network traffic.

The question was what function &quot;returns&quot; an inode.Those functions return a error code, you pass in a stat structure and the function populates that structure.He was saying (correctly), that they don&#x27;t return (in the classic C sense) the inode. They return an error code.To me that is a big difference...int lstat(const char path, struct stat buf);vs 
stat* lstat(const char *path);2 completely different functions.

But what if Google is down and you are tasked with diagnosing it?

Not agreeing with the recruiters attitude but &quot;if Google is down you will need to know this to diagnose what the problem is.&quot; was his comment. 
In that case the developer should have cheat-sheets printed on their desks.

He wouldn&#x27;t need to know it for the job, but as others have said he probably would know it if he really had the experience on his resume. That&#x27;s the value for me sometimes in questions like that, especially in an initial phone screen, to decipher if you&#x27;re bullshitting or exaggerating your credentials.

&quot;I knew all these answers too, because I was a developer in the 1990s.&quot;I think you&#x27;re onto something greater. Maybe Google should hire more smart people who were developing in the 1990&#x27;s as they&#x27;re a nice, middle ground between ambitious, young folks and the been-there-done-that greybeards. Them realizing such people already know all the shit they&#x27;re trying to teach their developers might be way to sneak older people into these tech firms. Haha.

Technically stat neither returns nor fills in the inode, because the actual structure of the inode can vary by filesystem, and the inode will contain more fields than struct stat.In a recruiting situation, if the recruiter is going all &quot;right and wrong&quot; on the interviewee, they should know these dirty details.I also disagree on the ethernet address length. You know how long IPv4 addresses are, you know how long IPv6 addresses are, why would it be so extraordinary to also know how long MAC adresses are?I think it could be useful to find out in an interview whether the applicant knows stuff because they have actually implemented low-level code and gained an intricate understanding, or whether they just did used some high level APIs and were never interested in more details than &quot;I have a handle right here, it does all I need&quot;. For some positions that would be an important distinction.Personally, if I were interviewing people, I&#x27;d hire the guy that explains something to me that I did not know, but that I find interesting and would have attempted to understand, too.

&gt; his bickering over stat() &quot;filling in&quot; rather than &quot;returning&quot; was bogus, for what it&#x27;s worthWhile I agree, I have seen interviewers where if he answered &quot;fstat&quot; they would have come back with &quot;Wrong! fstat passes the structure back by reference, it does not return it!&quot;With this style of interview I can&#x27;t blame him for thinking it might be a trick question and then trying to explain why he didn&#x27;t say those other functions as an answer.

I don&#x27;t think these questions are unreasonable as a spam filter. Yeah, they&#x27;re trivia, but if you had actually worked in that space, I&#x27;d be surprised if you didn&#x27;t know a lot of that trivia just by virtue of exposure.The rejoinder, of course, is that it&#x27;s probably misguided to structure your recruiting around a spam filter.

What&#x27;s bogus about his comment on stat() exactly. It literally returns an int and fills in the struct that it takes as a parameter. If he had no idea what it did he could check Google or the man pages which are going to give him some variety of the same answer such as the two examples below.Linux:<pre><code> int stat(const char *path, struct stat *buf);

 stat() stats the file pointed to by path and fills in buf.
</code></pre>
Mac OS X 10.11:<pre><code> int stat(const char *restrict path, struct stat *restrict buf);

 The stat() function obtains information about the file pointed to by path. 
 Read, write or execute permission of the named file is not
 required, but all directories listed in the path name leading to the file must be searchable.
</code></pre>
It may be pedantic but stat definitely doesn&#x27;t &quot;return&quot; an inode. One of the problems with technical interviews like this is you often have no idea what the interviewer is really looking for. Some might just want to know what you would use to get the information about an inode and another might be seeing how you describe it as a test of your knowledge of pointers or something. Often it&#x27;s impossible to know and that could have been easily a trick question where the right answer is &quot;there isn&#x27;t one but you can use stat to get this information as it takes a struct pointer and will place the information into that struct&quot;. Of course you know that better than most or you wouldn&#x27;t have made stockfigher.

I knew all these answers too, because I was a developer in the 1990s.There is absolutely no purpose to knowing off the top of your head how long an ethernet address is, or even what system call will retrieve an inode (his bickering over stat() &quot;filling in&quot; rather than &quot;returning&quot; was bogus, for what it&#x27;s worth). The top Google search result for each of these questions has the answer. Knowing these things isn&#x27;t part of being a practicing programmer; knowing how to find out is.

UNP is excellent! Haven&#x27;t opened it in a long time though.

And then, the world developed robust, well-maintained, open source libraries that do all of that for you. Now, in 99% of commercial software engineering, you don&#x27;t actually need to know any of those details, in the same sense that you no longer need to know exactly how many transistors and diodes are used to build the adder on the CPU. It simply no longer matters in almost all cases. Your time is much better spent moving on to higher level problems.If you&#x27;re interviewing for the rare 1% case where the job will actually require you to tinker with a TCP stack, then by all means ask about those details. For 99% of programming jobs today, it&#x27;s irrelevant trivia used as nothing more than an ego measuring device.

Yep, I was doing network programming back in the late 90&#x27;s and you pretty much had to have a good grasp of winsock and linux network calls and responses because the level of abstraction was much lower.That said I can&#x27;t recall any of that stuff mattering in the stuff I do for work for quite some time and most of it is a google search away.That interview was just weird, it was like asking the boiler pressure for a steam train to someone who was a master engineer working on electric trains.

The amazing part here is that so many people still remembers all those little details. I still know how it works, but honestly I forgot all the details long time ago.

Is TCP&#x2F;IP Illustrated still the gold standard of books on the subject? I read it fifteen or twenty years ago and would like to finally own a copy (I borrowed it from the library), but not if there is something better.

People that have been around know such things.   At one point, Richard Stevens (RIP) was god.  Every programmer had a copy of TCP&#x2F;IP illustrated, Advanced Unix Programming and Unix Network Programming.   If you wanted to do anything network, you had to write your own servers, you had to understand the details.  The breath of knowledge was wide and the depth was just as deep.  Ask around on HN, and you will probably be shocked how many that know such things. :-)

What are the components that make up a color?&quot;Well, it depends what color model you are using, which differentiates between additive and subtractive color mixing, the medium used (print, screen, etc.)...No. It&#x27;s Red, Green, and Blue.

To be entirely objective, I have run into recruiters of that variety in my career. However I&#x27;ve also had plenty of recruiters that were able to pick up on one&#x27;s deeper knowledge of a given concept and accept that as correct, even if it wasn&#x27;t the &quot;textbook answer&quot; they were looking for.

I had something this happen to me in an interview. Miserable experience.

I get it, but do doctors really know what the size of mitochondria are?  I can&#x27;t think of anything any further from the realm of their day to day.

Experienced this more times than I wish to admit.

Perfect analogy. And you just jogged a bad interview memory - &quot;thanks a lot.&quot;

Except that, for some reason, medical professionals don&#x27;t have to put up with this shit.  But we do.

Medical Recruiter: &quot;What are mitochondria?&quot;Applicant Doctor: &quot;They&#x27;re double membrane-bound organelles found in all eukaryotic organisms, commonly between 0.75 and 3μm in diameter, that generate most of the cell&#x27;s supply of adenosine triphosphate&quot;Recruiter: &quot;No. They are the powerhouse of the cell&quot;.

&quot;I mean who would know just how long a MAC address is? Or what the actual SYN&#x2F;ACK etc tcp flags are?&quot;This isn&#x27;t even my domain, but I remember some of this stuff. I had to dig in to this area years ago, and was knee-deep in that level for several months debugging and configuring stuff. Some of it stays with you, even if you&#x27;re not there any more.For someone applying for a director of engineering, I&#x27;m kind of split as to whether this should be required &quot;off the cuff&quot; knowledge. Would certainly help, but seems it would depend on the culture of the company - how hands-on they expect director-level folks to be - some companies seem to want that, some don&#x27;t.

&gt;...but I&#x27;m sure even a relatively junior coder would be able to tell if someone was just doing technical word salad.In interviews, I&#x27;ve had people give answers I&#x27;m not familiar with. The easy thing to do is to ask them to explain how their answer works. You get to vet whether it seems legit, and as a bonus see how good they are at communication in a genuine teaching moment.

Anyone who&#x27;s done any networking will know how long a mac addresses is

The right solution is to have a technical person conduct this test.

I&#x27;ve had questions that drilled down into what bytes are in a TCP&#x2F;IP packet on an interview. If you say you know about TCP&#x2F;IP on a resume, it&#x27;s fair game. FWIW I got all the bytes but the checksum; apparently no one ever remembers the checksum byte. :)What&#x27;s completely wrong with the situation is that the recruiter was saying &quot;wrong&quot; when he was giving detailed answers. The right solution is to have a non-skilled recruiter take careful notes, asking the candidate to repeat if necessary, in the cases where the candidate insists they know the answer.Then only in cases where the candidate can&#x27;t answer a question should it be marked as &quot;wrong.&quot; In other cases the answers should be run past someone knowledgeable.I&#x27;ve gone through a phone interview with Google myself, and it was nothing like this: I spoke with a real, skilled engineer, and there was nothing like a &quot;gotcha&quot; question where I had to guess the exact term he meant. Well, except where he asked &quot;I bet you know what my next question is&quot; and I didn&#x27;t guess &quot;How can the algorithm be faster?&quot; But he didn&#x27;t count that against me. :)

I&#x27;ve been responsible for writing network code on embedded systems, you do end up memorizing lots of the structures and constants from the specification because you&#x27;re constantly looking at them &quot;on-the-wire&quot; to make sure it&#x27;s write (or debug it if it&#x27;s wrong).

&gt; I&#x27;m amazed he knew things in such detail. I mean who would know just how long a MAC address is?Not surprised by the questions. Most are [somewhat] common and not particular challenging.I am surprised by the stupid counter-answers from the recruiter thought. This guy should not be giving phone interview.

That&#x27;s not the problem, a non-tech recruiter cannot assess the correctness. Even the simplification can be done, which I disagree, the answer will be rejected because it&#x27;s not a literal match. That, is the problem.

&gt; this guy always has to be right and would work terribly with people of different levels of skillsAre you talking about the recruiter?

Don&#x27;t think so, or would you consider the O(n) notation something a non technical user would want to follow?

I can&#x27;t fathom a scenario where a tech person would need to dump down something like SYN-ACKs and iNodes for a non-technical person. It&#x27;s one thing if you are trying to explain performance trade-offs to non-technical users or colleagues. But not lower level protocol details.

If this was the point of the test he should have been told this was the point of the test.  Gotcha&#x27;s make poor interview questions.

I&#x27;m just reaching here, but is there a chance at all that the test wasn&#x27;t really about whether or not he knew the correct answers but more that he knew the correct answers and was able to simplify them to the extent that a non-technical user could understand and compare them? I have a feeling that Google is far more interested in someone being able to get their point across than someone that just wants to sit there and argue about whether or not an answer is right. Just based on reading his responses, I got a condescending vibe and a vibe that this guy always has to be right and would work terribly with people of different levels of skills. At a Director-level position, that kind of skill is the most basic skill you need to have.

&quot;Diploma says DO, not MD. You don&#x27;t get the job.&quot;<a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Comparison_of_MD_and_DO_in_the_United_States" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Comparison_of_MD_and_DO_in_the...</a>

&gt; I wonder what would happen if ordinary people recruited for medical doctor jobs? Would you be comfortable rejecting a guy who&#x27;d been in medical school for 10 years based on his not knowing what the &quot;funny bone&quot; is? Wouldn&#x27;t you tell your boss that you felt a bit out of that league? It&#x27;s amazing you can get someone to do this without them going red in the face.The medical recruiter asks to see the diploma. The doctor shows his degree and gets the job.

If you&#x27;ve had to troubleshoot network corruption at scale, you memorize packet details real fast :)

&gt; I&#x27;m a student (not someone with 10 years of experience) but I knew that. 3 bytes for the vendor block, 3 bytes for the device.That is why you know the answer. Come back in 10 years and let us know if you still know it. What you think might be mainstream in a computer science class are rarely used in application. And if they are they can be easily looked up.I used to know the exact effective distance of a CATV cable when I was a student. Useful? Sure. Something I need to remember for the rest of my life? Definitely not.As another example, in 17+ years writing ISO level 7 programs, I have never once needed to use the Mac address.

&gt; Uh, my main languages are PHP and Python (high level stuff) and I&#x27;m a student (not someone with 10 years of experience) but I knew that. 3 bytes for the vendor block, 3 bytes for the device.Thanks for the refresher! I knew that too when I was in college. Good luck remembering that 5 years from now :)

Yeah, you&#x27;re a student; you probably just took a class on this stuff. You&#x27;ll forget it soon enough if you don&#x27;t use it.

&gt; I mean who would know just how long a MAC address is?Uh, my main languages are PHP and Python (high level stuff) and I&#x27;m a student (not someone with 10 years of experience) but I knew that. 3 bytes for the vendor block, 3 bytes for the device.&gt; Or what the actual SYN&#x2F;ACK etc tcp flags are?Yeah the actual bytes, who ever uses that? A MAC address I&#x27;ve seen plenty of times in my life as hex, and I&#x27;ve seen the TCP setup flags being exchanged plenty of times in Wireshark and looked up the hex once when I was implementing TCP from scratch, but I still wouldn&#x27;t expect anyone to know that.&gt; You just need to know what they&#x27;re used forAgreed on that. I wouldn&#x27;t blame anyone for not knowing the size of a MAC address, I just didn&#x27;t think that one is that obscure.

A civil engineer is a professional engineer. I would assume your father is a certified PE [1].This rubbish flies in our industry since we are not professional engineers.[1]: <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Principles_and_Practice_of_Engineering_Examination" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Principles_and_Practice_of_Eng...</a>

My father is a civil engineer and late in his career, he interviewed for a senior position at a construction company. He figured out that the interviewer didn&#x27;t know what he was talking about early on the interview and confronted him with that. The interview came to a end, he was put in touch with a senior engineer at the company. He didn&#x27;t take the job for other reasons. I don&#x27;t think this kind of rubbish would fly in any other industry.

It&#x27;s true this is probably not exactly what happened, but Google does certainly ask these questions in phone screens and the interviewers read the correct response off of a cheatsheet.The fact that several Google people posting here literally can&#x27;t believe this is true shows how fucked up their process is.

I&#x27;m amazed people actually believe the account is accurate. The guy failed the interview, and goes to write a blog post to lash out at google (and try to mess up future interviews for them), So of course he write it to make him look good, and the questioner bad.We will stop here because it&#x27;s obvious that you don&#x27;t have the necessary skills to write or review network applications. You should learn the Linux function calls, how the TCP&#x2F;IP stack works, and what big-O means to eventually qualify if you are interviewed at a later time.Come on Hacker News, I know you hate recruiters, but do you really think that happened?

&quot;I&#x27;m amazed he knew things in such detail. I mean who would know just how long a MAC address is? Or what the actual SYN&#x2F;ACK etc tcp flags are?&quot;I was more impressed by how simply he explained many of them in real-world terms. Especially countering Big O. That plus what he&#x27;s already built indicates he&#x27;s exactly the kind of person to direct the sort of projects they&#x27;re building. That they filtered him with this garbage Q&amp;A speaks volumes about Google&#x27;s ineffective hiring. Plus, indicates what kind of people might have made it into the organization.

Actually, most people with a networking background would know this kind of detail (and more). It&#x27;s the relevance (and appraisal of answers) that is shocking.

Oh hey what are you going to do with your million?

You proved P = NP??? In a little more than 45 mins or did you misunderstand the question?

P = NP if N = 1, or P is 0.  That doesn&#x27;t take 45 minutes.

Maybe they didn&#x27;t hire you because your claim that you proved P=NP in 45 minutes is ridiculous.

1) His level of knowledge is reasonable for someone who does networking programming full time2) We are getting description from him. When I interviewed at Google, they asked me to prove that P is equal to NP, and I did, but they said that it took me longer than allotted 45 minutes and I didn&#x27;t get an offer.

&quot;who would know just how long a MAC address is?&quot;I&#x27;ve been looking at them on NICs, adding them to DHCP databases and using them for simple authorizations in iptables (and prior equivalents) since, oh, 1992 or thereabouts. They weren&#x27;t new then.It would take me a moment to count them in my head, that&#x27;s all. &quot;aa, bb, cc, dd, ee, ff -- that&#x27;s six bytes&quot;.

&gt; ... like they can memorize but not think.Which is really bad, when they&#x27;re trying to hire people for a job where thinking is the fundamental skill...

I was interviewed once by someone who would cut me off after just a sentence or two. He asked a fairly open question like, &quot;Tell me about hash tables.&quot; I knew I should have just said, &quot;Lookups are constant-time,&quot; but I started off with a description of their behavior instead, and sure enough he cut me off and said, &quot;Lookups are constant-time.&quot; Some people are just not very good interviewers, but also some tech people&#x27;s rules are too iron, with no exceptions, like they can memorize but not think. They say interviews are a two-way communication, and that one certainly was. :-)

Same things I thought. His answers are amazing!

It&#x27;s actually pretty sad. It seems to weed out the smartest folks so that you only hire those mediocre enough to not understand these problems within the hiring test.

I wouldn&#x27;t be surprised if the interview process for experienced engineers is unrefined.
At this point in Google&#x27;s lifecycle most of the qualified, experienced engineers who would want to work at Google already do. There are orders of magnitude more new-grad engineers to interview and so it makes sense that they would lack the practice and refinement on those candidates, even if those are the most valuable candidates to hire. The recruiter might have just been confused and gave him the standard list for any &quot;technical&quot; job that they have to use for new-grads and so cannot ask things with nuance.When I was an interviewer at Google it felt like 90%+ of interviews were with candidates who had less than four years of experience. Probably half were fresh out of college. After the fifth candidate in a row who can&#x27;t do simple recursion or algorithmic analysis (and I mean simple) you get pretty discouraged. In one phone interview I got to interview an experienced engineer with over twenty years of experience in C. He completed the question I usually have to spend 45 minutes on with a new-grad in &lt;10 minutes. It was probably my favorite interview of all time because I actually got to discuss the subtleties and he reaffirmed that I could maintain high standards.

It makes me think there&#x27;s a screening recruiter, non-technical, before the &quot;real&quot; interview process. At big companies like Google with so many people interviewing, I can imagine these kinds of kick out questions.Still, it shows a massive company trying to streamline some things and failing terribly. I personally wouldn&#x27;t want to work at Google today. What might have been cool once is now nothing but a standard large company like IBM or Amazon. There&#x27;s a great Quora post by some former Google people that say as much.

Personally, these questions and the way the recruiter asked them reaffirm my view that Google would not be a place I&#x27;d like to work at as a experienced software developer. First the recruiter&#x27;s lack of technical knowledge points toward a beaurocratic or management first mindset common (necessary?) in such a large company. Second, the questions and expected answers seem biased toward just graduating but smart engineers who don&#x27;t actually have experience to realize the subtleties present in complex systems. The recent lawsuits regarding age discrimination reinforce the notion that this interview setup is biased (intentional or not) toward inexperienced and thus likely younger applicants. Perhaps that&#x27;s not a bad thing per se as Google&#x27;s corporate development style likely would handle inexperienced but smart developers who can &quot;mold&quot; into the system better. Or it could be more nefarious as to lower salaries by getting younger devs. In reality, probably a mix of both.

I think the interviewee after realizing the shortcomings of the tester&#x27;s understanding and the author of said &quot;right answers&quot; should have interrupted or ended the probing after a few red flags and politely suggested something along the following lines...&quot;Look I get the gist of what this screening is trying to do but please write down my answers and feed them to the technical staff who authored this Q&amp;A and have them verify their misunderstandings and have them get back to me please. I am very confident you all have subtle details mixed up and could benefit by a strong candidate who understands such details with a higher level of fidelity.&quot;Guess its another loss given to the recruitment focus on avoiding false positives over false negatives.

&gt; I mean who would know just how long a MAC address is? Or what the actual SYN&#x2F;ACK etc tcp flags are? You just need to know what they&#x27;re used forDepends on what you actually do in your job, I suppose. If you use that information regularly or troubleshoot networks regularly you would soon enough memorize it.&gt; and if you need the specifics, you&#x27;ll find out with a single search.Someone doing network programming _does_ need the specifics.

Understanding SYN&#x2F;ACK, and what they are used for, can be quite useful in troubleshooting networking issues so I can see why this would be useful in a SRE or other role that dealt heavily with communication.

It happens because great developers do not do recruiters&#x27; job :)

&gt;Gwan is less known around because it&#x27;s not OSS but I believe still achieves unmatched performance.ERR_CONNECTION_REFUSED

Gwan&#x27;s performance is the real deal. It&#x27;s also the least stable piece of software I have ever used in my entire life.That last sentence is important.

He&#x27;s one of the best programmers you can get (in the world) when it comes to networking and friends. Gwan is less known around because it&#x27;s not OSS but I believe still achieves unmatched performance.This interview must be some kind of a mistake. It&#x27;s like testing Richard Feynman&#x27;s knowledge with questions about multiplication table (where you have wrong answers on your table).It&#x27;s so stupid that it&#x27;s almost not possible to laugh.

Who would know just how long a MAC address is? Jeez man, these things are literally written as 6 bytes separated by colons. Asking a network engineer this question is about like asking a postman how long a zipcode is.

I would have failed the question.  I don&#x27;t remember how many bytes in a MAC address.  That&#x27;s something that can be easily looked up.

&gt; I&#x27;m amazed he knew things in such detail. I mean who would know just how long a MAC address is? Or what the actual SYN&#x2F;ACK etc tcp flags are?It&#x27;s what happens when you get old. There were commonly C compilers on 8, 16, 20, 24 and 32 bit architectures, some big-endian, some not, and you couldn&#x27;t google things in 20 seconds back in those bad old days when old farts were building the networks that would be used to connect to us all to google.(And he has apparently been in networking since)

&gt; I mean who would know just how long a MAC address is?Hmm, let me think. Somebody working with routers for 10+ years.&gt; ..what the actual SYN&#x2F;ACK etc tcp flags are?Same.I mean, these things are just as familiar to me as grade school multiplication, by now..I bet there&#x27;s things familiar to you which I couldn&#x27;t even be bothered to memorise and instead had to turn to a search engine for.

I&#x27;m amazed he knew things in such detail. I mean who would know just how long a MAC address is? Or what the actual SYN&#x2F;ACK etc tcp flags are? You just need to know what they&#x27;re used for, and if you need the specifics, you&#x27;ll find out with a single search. He seemed to know that as well though. Kernighan for bit twiddling algos, that kind of thing.It&#x27;s a bit strange to have someone non-technical interviewing a techie. You end up with stupid discussions like the one about Quicksort. If you point out qs is one of several things with the same big-O, you&#x27;ll probably also get it &quot;wrong&quot;. But the real problem is that a guy who is just reading off a sheet can&#x27;t give any form of nuanced feedback. Was the guy blagging the sort algo question? Did he know if in detail? Does he know what the current state of research on that area is? There&#x27;s no way to know that if your guy is just a recruiter, but I&#x27;m sure even a relatively junior coder would be able to tell if someone was just doing technical word salad.I wonder what would happen if ordinary people recruited for medical doctor jobs? Would you be comfortable rejecting a guy who&#x27;d been in medical school for 10 years based on his not knowing what the &quot;funny bone&quot; is? Wouldn&#x27;t you tell your boss that you felt a bit out of that league? It&#x27;s amazing you can get someone to do this without them going red in the face.

It was not the answers but the cultural fit ehehe...

Having had some experience with big companies, I doubt that. They&#x27;ll probably just write this guy off as being &#x27;non googley&#x27; or whatever, and keep on keeping on.(assuming the interview really went down this way. Which I have a hard time believing)

Doubt it.This is a pretty standard google first interview. Nothing out of the ordinary.The recruiter will not suffer for following standard proceeding.

I feel sudden brick wall awaits in a recruiter career path once the google people read that.Not only because of that interview, but the questions transforms into - who else have we missed from the great guys ...

Are you surprised though? If a person who does not have the merit to judge the answers very well this is what happens.

Why the hell is a non-technical recruiter giving these tests? Google engineers or managers should be administering it, not the high school dropouts who make up the recruiting industry. There&#x27;s so much wiggle room here that have a rote-memorization style test for engineering is completely crazy. I&#x27;ve never had a technical test given to me by a recruiting firm. Its always a call with someone from the company, if not the hiring manager.Perhaps this is what they give to &#x27;second string&#x27; applications to make HR happy while the guys who actually get the jobs are friends of the hiring manager or team leads. Or its a H1B ploy to say, &quot;See Obama, we need tech talent. Look how terrible our domestic talent is. They score 40% on our tests!&quot; Meanwhile Bombay Upstairs University has a wink-wink-nudge-nudge deal with Google hiring managers who accidentally leak the test on a &#x27;forgotten&#x27; ftp site.Everything about this is fishy. I think there&#x27;s fraud here, not just incompetence. I&#x27;ve been the hiring stooge for shops who have already made their decisions before and its always terrible and, frankly, hurtful. These are the signs of a non-serious &#x27;stooge&#x27; interview.

This is actually a real danger in hiring: personal bias.

I feel like someone that&#x27;s not unusually unintelligent should have been able to make the connections between SYN and synchronize or ACK and acknowledge. I don&#x27;t know why people are defending the recruiter. The process is dumb, but so is an interviewer that asserts himself as an authority when he has to read from a sheet of paper. I don&#x27;t expect him to be a technical expert, but I do expect him to know when he needs to consult a technical expert about a particular candidate&#x27;s answers. If this is an accurate representation of his interviewing skills, he needs a new job.

In my experience Google recruiters do have a sheet with alternate answers, but those can&#x27;t cover everything.

Surprising they wouldn&#x27;t have given the recruiter a list of alternate acceptable answers.Almost feels like the recruiter was conspiring against you, but why?

Ha ha re:&gt; 10. what is the type of the packets exchanged to establish a TCP connection?&gt; Me: in hexadecimal: 0x02, 0x12, 0x10 – literally &quot;synchronize&quot; and &quot;acknowledge&quot;.&gt; Recruiter: wrong, it&#x27;s SYN, SYN-ACK and ACK; if Google is down you will need to know this to diagnose what the problem is. We will stop here because it&#x27;s obvious that you don&#x27;t have the necessary skills to write or review network applications. You should learn the Linux function calls, how the TCP&#x2F;IP stack works, and what big-O means to eventually qualify if you are interviewed at a later time. Good luck, bye.That&#x27;s embarrassing.

I know, that&#x27;s why I would be surprised if this is standard practice for high level positions. But even if it is, this is unusually bad execution. My recruiter had at least surface level knowledge of the concepts and was able to ask follow-up questions if I got something half right, he wasn&#x27;t just looking for verbatim responses.

&gt; It&#x27;s just so they can avoid setting up a real interview with someone who doesn&#x27;t know the basics.They&#x27;re (allegedly) interviewing for a Director of Engineering. If google can&#x27;t even google the guy and look at his track record, code, whatever, they don&#x27;t deserve to have competent people working for them.

&quot;Is this common for higher level positions at Google?&quot;No, and in fact, it&#x27;s so far outside the norm i&#x27;m not even sure what to make of it.
Like I said elsewhere, my best guess is that he was really being evaluated for a much lower level TL&#x2F;M position in SRE or somewhere.(The detailed linux questions are usually a giveaway that SRE is involved)

Is this common for higher level positions at Google? I&#x27;ve gotten a similar kind of phone screen, but for an entry level software engineer position, and getting a few &quot;wrong&quot; was not a big deal. It&#x27;s just so they can avoid setting up a real interview with someone who doesn&#x27;t know the basics.If the call actually went like this, it seems like you just hit a new&#x2F;not very good recruiter.

I would have stopped you half way through. You obviously know enough to move on.

&gt; This reminds me of the infamous Nigerian emails. They are so obvious for a reason - avoiding false positivesUnrelated to the main topic but I wonder if email providers could effectively fight those scammers by having chatbots automatically reply to their emails. That would make a bunch of false positive to sort through and possibly make the whole thing uneconomical for the scammers.

This reminds me of the infamous Nigerian emails. They are so obvious for a reason - avoiding false positives - [1]. Cannot fit this same scheme in a Senior position recruitment at Google though.[1] <a href="http:&#x2F;&#x2F;uk.businessinsider.com&#x2F;why-nigerian-scam-emails-are-obvious-2014-5?r=US&amp;IR=T" rel="nofollow">http:&#x2F;&#x2F;uk.businessinsider.com&#x2F;why-nigerian-scam-emails-are-o...</a>

Are they hiring a reference manual or a director of engineering?

Having been through several goog interviews I find it hard to believe that the interviewer would dismiss his highly technical explanations&#x2F;rebuttals in that way. Seems more likely this is his interpretation of the tone of the interview, rather than a transcript.

Why not? Especially in a managerial position, you will occasionally need to build a rapport with people who are convinced they have good technical knowledge, but don&#x27;t. Being able to talk to them productively instead of picking fights is a skill worth selecting for.If you can&#x27;t figure out that the first person who&#x27;s interviewing you has answers on a sheet of paper and you&#x27;re supposed to parrot them until you get to the second person, how are you ever going to figure out that the first person you&#x27;re selling to has some business requirements on a sheet of paper and you&#x27;ll never get to the second person until you parrot those?&quot;Oh, we&#x27;re not actually using Docker, we&#x27;re using rkt, which is a compatible reimplementation of --&quot; &quot;I&#x27;m sorry, I&#x27;ve been told Docker is a requirement. We can&#x27;t use your Cuber Netty thing until you support it. Bye!&quot;

Yah I get the feeling you&#x27;re supposed to &quot;act stupid&quot; so you can get to the next interview.Definitely not an idea way of getting candidates - you are selecting for people who know how to manipulate screens (and thus increasing the risk of getting a bad candidate), or rejecting otherwise knowledgeable people who just don&#x27;t have the time&#x2F;inclination&#x2F;&quot;social savvy&quot; to pretend to be stupid.

At that age, it becomes way more satisfying to educate somebody, rather than try to force your way in.

While this recruiter probably failed at their task, I do agree that there is skill in communicating technical information to non-technical people, and it is a valuable skill. I wouldn&#x27;t be surprised if the author of this blog post acted a bit curt or acrimonious to the recruiter, and stubbornly insisted on being &quot;right&quot; rather than communicative and agreeable.Know your audience.

You mean the wrong answers. Great, in a technical interview for the Director of Engineering (not legal, not HR) Google expects you to know when to lie and give the wrong technical answers and when to give the right technical answers because this is essential to the Director of Engineering position? What are they hiring someone to be complicit in their next wage suppression scheme?

Doesn&#x27;t make sense given the end of the interview, unless the recruiter was being completely disingenuous. The recruiter recomended the interviewee getting some education in the topics he asked about.

The meta-question is, &quot;Do you have the social savvy to give the conventional answers when being judged by someone who doesn&#x27;t know what he is taking about?&quot;

Most people, at every level, are terrible at interviewing and hiring, and in fact a surprising number are conscious of that fact. As a candidate, you just have to know this going in, and expect to be made to jump through a lot of irrelevant&#x2F;stupid hoops. Eventually you find someone who is adequately amused by your performance to give you a chance and hire you.I don&#x27;t think the recruitment industry is going anywhere anytime soon. I just wish there were some worthwhile recruiters in it (out of the dozens and dozens of recruiters I&#x27;ve dealt with, I&#x27;ve met one who is semi-competent (which is to say, not fully competent)). This is an area I&#x27;d love to try to startup in, I don&#x27;t think any of the startups I know about are doing it right.

It&#x27;s a freaking psychological profile test dressed up as &quot;tech screen&quot;. Your friend&#x27;s personality was effectively filtered.IMO, the very first step in entering googleplex&#x2F;facebook is accepting that it is OK to lend your talents to build a global surveillance &amp; propaganda platform. These tests seem to be probing for other desired characteristics as the next step.

A friend of mine was offered a similar position and had a similar interview. Annoyed by the questions and the &quot;my solution is the only right one because I have to stick to this script&quot;-attitude he asked: &quot;So what do you want me to do at Google, how can I help you&quot; the recruiter could not really answer and ended the call shortly after.

You can fire this Director of Engineering.

That is nearly the answer i give to every questions. But when i asked for more detail no one likes it. They all simply want &quot;answers&quot;.Well much like most things in life, there is no such thing as concrete answer.

A brilliant professor of mine liked to say &quot;The Answer of the Engineer: it depends&quot;.

It is scary if this is how Google hires engineering directors. Passing this test would select all the bad candidates! Answers to many technical solutions are &quot;it depends&quot;. If you have a director that &quot;knows&quot; the right answer, they won&#x27;t be very likely to encourage their engineers to experiment and find the best solution to the specific problem at hand.

This is what happens when a non-technical person conducts a technical interview.Though I&#x27;m surprised this happened for a Google interview. Their recruitment workflow should now have allowed this to occur.

I&#x27;ve never interviewed for director-level positions, but I have definitely received a few technical questions from recruiters who didn&#x27;t seem technical. The questions were generally easy.I still think it&#x27;s a bad practice, though. If you absolutely insist on doing so, multiple choice is the right way to go.

This wasn&#x27;t a technical interview. This was a phone screen. As dumb as it is, having non-technical recruiters do a &quot;technical&quot; phone screen like this is increasingly common. I don&#x27;t think this practice is completely justifiable, but part of the root cause is the enormous volume of grossly underqualified people who apply for any particular position. It takes a hiring company a large amount of work to vet a candidate but it takes a candidate nearly 0 effort to apply for a job.

It&#x27;s standard practice at the Googles and Facebooks of the world to have a recruiter lob a bunch of canned questions at you during a phone screen, including Director-level phone screens.  My experience was a lot different that the writer&#x27;s, however.  The recruiter was friendly and I was friendly back.  I think I missed one or two but still went on to the next round.  This guy comes across sounding like a pompous asshole and I wouldn&#x27;t want him as my Director of Engineering, either.

It saddens me slightly how far down the page your comment is, as that&#x27;s exactly what&#x27;s going on here. I&#x27;ve also (and also not at Google) had a hand in constructing these. It&#x27;s a nightmare on the hiring side, too.We tried very hard to tailor our questions such that the candidate — if they new the answer — would only give one exact answer. Ours included a download of sample data to operate on; one was a question along the lines of &quot;count the number of lines in any .h files that contain the following pattern &lt;human description of the patter&gt;&quot;; the answer ends up being a single integer.My experience has been — overwhelmingly — that recruiters hiring for technical positions are incredibly non-technical. This wouldn&#x27;t necessarily be a bad thing, except for, as mentioned, the huge amount of winnowing required to just find candidate that are suitable to advance to a phone screen with an engineer.&gt; There are a LOT of people out there who grossly inflate their competency.Exactly this.

Many companies do a pre-interview screen performed by recruiters to filter obviously bad candidates. I have personally helped construct such screens -- though not at google.Ensuring the recruiter understands the difference between an answer over their head and a wrong answer is crucial. What we did was to make sure the recruiters understood that answers they didn&#x27;t understand needed to be surfaced to someone in engineering for a sanity check.Often our strongest candidates would give an answer which recruiting wasn&#x27;t capable of vetting. The key is making everyone aware that the answers may not fit the script, and that this is OK. Recruiters shouldn&#x27;t be judging the technical details of an answer; they should be looking for &quot;gosh I don&#x27;t know I&#x27;ve never seen that.&quot;Pre-screening is sadly necessary when doing high volume recruiting. There are a LOT of people out there who grossly inflate their competency.

It&#x27;s just a phone screen. I got some of these questions from a Google recruiter on a phone screen, for an SRE manager position.  Actually the recruiter asked me if I prefer questions focusing on system or on complexity and general CS topics, and I chose the latter. After passing it I got to a phone interview from an engineer, which had more questions and some coding. But this sounds about right for an initial screening.

What the 5 different subjects of those 60 minute interviews?

As stated elsewhere, these questions have been very typical for what I&#x27;d call &#x27;first level phone screen&#x27; for SRE positions at Google.I interview with Google every couple of years for fun; I always learn a lot and have a lot of fun with the on-site interviews.The recruiter does indeed ask these kinds of questions. I&#x27;ve been asked most of the ones mentioned here, some of the multiple times.I find it extremely hard to believe that the answers the recruiter had in front of them was wrong.So to be clear, how it has worked every time I&#x27;ve done the dance with Google is:1. An initial, non-technical recruiter chat.2. The recruiter gives me a series of technical questions with extremely clear-cut answers, as described in the article.3. A phone screen with someone technical.4. A second technical phone screen.5. On-site, 5-6 50 minute interviews.

This is very hard to believe. In most big tech companies technical interviews are conducted by technical people, not recruiters. Especially for a senior role like Director of Eng. the experience described here stretches credulity.Also most of these companies have policies which don&#x27;t allow any feedback to be given on interview performance. In light of that the recruiter saying &quot;you don&#x27;t have necessary skills&quot; is extremely surprising.

&quot;Order&quot; would be the commonly used term, right? At least in my native language.

Big-O technically means &quot;worst case&quot;, which for Quicksort is O(n^2), although usually (in normal cases) it runs better than that.  Other sort algorithms can guarantee not worse than O(n log n).

I assume they want the average O(n log n), but who the heck knows given the script from above.I get the feeling this: <a href="http:&#x2F;&#x2F;bigocheatsheet.com" rel="nofollow">http:&#x2F;&#x2F;bigocheatsheet.com</a> would come in handy

What the heck is a &quot;Quicksort big-O score&quot;? I&#x27;ve never heard anyone use the noun score when talking about complexity analysis.

Thing is, mortals like I know very little about workspace psychology compared to large scale operations such as Google. I have an inkling of an idea because I was repeatedly subjected to the same problematic environment too many times in my 10 year career as a simple software developer with no lead titles.Off the top of my head:There is always someone incompetent getting paid more than you. I think this means they are willing to spend money on someone else so you work harder to reach his&#x2F;her status like a horse going after a carrot it will never reach.There is always some kind of boss who will not indulge in technical details in order to understand how long a project would really take before pitching the idea to a customer. This is their way of &quot;getting things done&quot; or &quot;making hard things a reality&quot;.

Sorry, you got too many right, you&#x27;re not right for this role!Actually, come to think of it, I did have an interview experience like that, just a few years ago. At the time, my title was VP of engineering, but for a small startup. A bigger company was interested in hiring me for a VP-level (or maybe director level, I can&#x27;t remember) position, and phone screened me. After about ten minutes of talking, he said, &quot;are you sure you wouldn&#x27;t be more interested in our technical architecture team? You sound like a good fit for our technical architecture team&quot;.

Perhaps they are looking for a pointy-haired-boss type rather than someone who actually knows his thing.

Hehehe - that&#x27;s really funny. Why did they even need a &quot;body&quot; behind the test? If all the person is doing is reading from script - it could have been just web based.

Here&#x27;s the good news: Google will get the directors of engineering they deserve.

&gt; they should have skipped you straight to an phone interview with an engineer.When a family member of mine used to be an admissions officer at a small tech school, &quot;unsuitable&quot; (mentally ill, ppl who seemed like they would never get in, be able to pay or complain and tarnish the reputation of the school) used to come in often. They were given a screening test which was waived for all &quot;normies.&quot;This is how power works. It is not fair. With communication media like the internet, we&#x27;re starting to perceive it. The real challenge will be re-architecting our society to be more fair and clean up from the aftermath of our unfairness.

As a person who&#x27;s taken this exact same test:
1) it&#x27;s a pre-screen from an in-house recruiter. If you have 37 years of experience with computers and 24 years professional experience and any references&#x2F;reputation, they should have skipped you straight to an phone interview with an engineer.
2) Were you recording the interview? Because if it&#x27;s an accurate transcript the recruiter should be fired.I assume question 7 was actually &quot;give the Big-O of Quicksort&quot;, if you refused the recruiter might have assumed you didn&#x27;t know it and were trying to BS through the interview. If the recruiter was too stupid to ask, once that person is fired you&#x27;ll get a new recruiter. Google is (was?) notorious for contacting the same candidates every six months regardless of previous interviews.

This seems like a pre-screen where the recruiter is trying to assess the viability of a candidate themselves. But I would assume that based on experience with recruiters elsewhere, no idea if Google has a higher bar for their recruiters.

I know someone working in that area--apparently lots of string parsing.  It&#x27;s vitally important (moving money usually is) but not very exciting&#x2F;cutting edge (moving money usually isn&#x27;t--after all, it&#x27;s an old technology).

Wait. You didn&#x27;t know what position and location you were being interviewed for? That kind of arrogance from a company can only exist if the job seekers are desperate or perceive any job at Google as nirvana.

I only have a sample size of one, but I blew threw a phone interview with Google about a year ago -- and the guy giving it was very technically skilled. We talked about the subtler aspects of the questions as we went.They offered an in-person interview, of course, but I declined when I found out that they were only hiring for Google Payments in Boulder. I&#x27;m sure the job has interesting aspects, and maybe my imagination just isn&#x27;t up to the task, but I have a hard time figuring out how I wouldn&#x27;t go insane with boredom working on a system that just moves money around, not to mention frustration from working with the extra regulatory&#x2F;process restrictions that must be in place to keep compliance up...

Likewise, identical situation, identical thoughts&#x2F;response.I want to follow up on this and see what the deal is. It just strikes me as fundamentally wrong.I&#x27;m honestly hoping this is not actually a Google recruiter doing this; if it is, that&#x27;s just broken.

Google recruiters ask these exact questions in phone screens.

This strikes me as an interview with a recruiter. This is almost definitely not a legitimate Google interview.Maybe it was for a vendor&#x2F;contractor position.

I had exactly the same kind of quiz at the beginning of the year on a Google phone screen interview. Some questions were exactly the same. I passed apparently but declined the in person interview. The recruiter over the phone was way more technical though and accepted answers not strictly matching his response sheet (and was able to discuss the technical whereabouts)

I&#x27;ve had similar phone screens previously both at Google and Facebook. The depth&#x2F;specifics of some of the questions seems maybe a little deeper but it&#x27;s not too far off what I remember from my screens to be honest.

To add to the chorus: I had an extremely similar interview (including a few verbatim questions) during my phone screen for an SRE role at Google five years ago.

These are pretty much SWE&#x2F;SRE questions, not director questions.

My company posts positions available in a common area as part of an H1B related initiative in order to verify that no US techie wants the job before they go looking for the candidate.  They were offering around .6 of typical salary given position and location.  Shameful what is happening in tech right now, absolutely shameful.

I know some companies do this, but this is Google. There&#x27;s no incentive for them to hire H1-B&#x27;s if a equally qualified American citizen is available, since they are going to pay equal salary.

Google typically recruits people from Europe in European offices, and the same goes for offices in Asia, etc. The H-1B process is so painful that they do everything to avoid it, and only resort to it if they cannot find the candidates they are looking for in the US (which happens often on the scale of a company as big as Google).Some people responded saying that they might still do this to get underpaid employees from abroad... which is just silly. The salaries are exactly the same whatever your country of origin, and companies like Google are not the ones you should attack if you want to make a point about H-1B abuse.

The only way this makes &quot;sense&quot; is if you already have the candidate (or pool of H1-B candidates) you want in mind, but have to prove you opened up the position to the general public first.

As another data point, when I interviewed at Google a few years ago I had a phone interview with some of these exact questions.Also, judging by the smugness of the people I interviewed, the &quot;We&#x27;ll stop here because it&#x27;s obvious you don&#x27;t have the skills...&quot; part doesn&#x27;t surprise me either.

I was asked similar questions when I was hired as an SRE nine years ago. The recruiter stopped after a few correct answers, iirc. There was probably at least one I didn&#x27;t know &#x2F; guessed wrong on, but I don&#x27;t remember. They also asked me to rank my knowledge in several areas on a scale of 1-10, and I think the questions focused on areas in which I&#x27;d chosen higher grades. Following conversations with engineers also focused on those areas.As I understand it, this is meant to be a shibboleth a non-technical recruiter can use to spot an experienced software engineer &#x2F; sysadmin in a quick conversation (&quot;pre-screen&quot;). That&#x27;s a hard thing to pull off. They can&#x27;t ask someone to design a system, diagnose a problem, or write code because they&#x27;re unqualified to grade the answer. Instead, they ask some simple canned questions. The questions may not test essential, first principles sorts of knowledge, but if someone can&#x27;t answer any of them it&#x27;s a bad sign. The questions should have a small family of correct answers that recruiters can recognize, and the recruiter should just see that a candidate can get some of them right before scheduling a phone screen with a Google engineer. If the transcript is accurate, this process failed.

I&#x27;ve been at Google for five years as a SWE and I&#x27;ve been interviewing for 3 of those. I&#x27;d fail this pop quiz.This strikes me as bizarre and inconsistent with all the practices I&#x27;m aware of. The idea that we&#x27;d ask anyone this stuff, let alone director candidates, strains belief.

wow, thats especially terrible. The engineering interview pre-screen I had at Amazon was better than that. Can&#x27;t imagine why a company would do this for ANY technical position, much less such a senior one.I did have an HR pre-screen one time where they sent me a SQL literacy test (did not specify sql dialect). One of the multiple choice questions had all correct answers, in different sql dialects, with no &quot;all of the above&quot; option. I corrected the test. I did not get a call back.

If this is true, then it&#x27;s not surprising as usually the person asking the question is not up to speed on the details of the question being asked.I recall a question, which was how do you stop a socket from overflowing?I said what do you mean? If you write to a socket and it fails to accept the write it will fail. Sockets do not have any option to stop this.Naturally I failed the question yet I asked for the correct answer and was told the answer is simple, just use a cache.Nice answer, excepts sockets do not have caches.Correct Answer: Add a cache to the socket to stop them failing but that was not the question.While that answer is obvious, it was not the question asked!!

Google and all the other mega-tech corps do a similar tech phone screen where they&#x27;re reading off an answer sheet. (powers of 2, sort these functions, linux kernel calls, etc).Sounds like you dodged a bullet, this sort of thing speaks volumes to how the company is run internally and how you&#x27;d be treated there.

Perhaps, as of yet, Google isn&#x27;t living fully by their own principles (&quot;automating themselves out of a job&quot;), maybe because the tech isn&#x27;t quite there yet or maybe for other reasons. It remains to be seen if the ultimate solution to bureaucracy is AI&#x2F;ML or if it will just be a further extension of these same mechanisms.The problem with simple &quot;skills tests&quot; is they lack depth, and of necessity due to time, always will. The problem with performance tests is the opposite: they may take too long to reach a conclusion in human terms, during which time significant money is lost.Recruiters seem like a great candidate for replacement by AI. [edit: and I should add, most middle managers too]

I&#x27;m reminded of a story about Einstein in the 1930s. The fad at the time was to have a &quot;skills test&quot; which was supposed to place you perfectly according to your results.Einstein took the test, and did terribly. Everyone was shocked, and asked him what was going on. His response was &quot;The questions tested memorization. Why would I remember what the capital of North Dakota is? I can look that up in a book!&quot;Perhaps unsurprisingly, the tests faded out of popularity.It&#x27;s sad that 100 years later, even Google hasn&#x27;t learned that lesson.

I&#x27;m feeling a bit confident looking at the questions. Woo.

This sounds like a phone screen--i&#x27;ve never signed an NDA for phone screens or at any stage of the interview process for that matter.

Pretty sure Google&#x27;s interview questions are under NDA.. Brave!

Completely agree with you. Seems more like a personality mismatch, than a lack of technical skills.Seems like both parties were at fault here:- The interviewee for being a little brusque about his answers, and despite understanding the _intention_ of the question, giving potentially off-putting, difficult answers to check over a quick phone call.- Google recruiter for seeming to ask questions she&#x2F;he was a little out of her&#x2F;his league to ask. If you don&#x27;t understand the nuances in and around the questions you&#x27;re asking, you are not qualified to evaluate someone&#x27;s responses.Wouldn&#x27;t worry about any one particular interview, there are too many great opportunities in tech right now to let any one bad interview let you down!

I can see both sides here. Google does want to make sure even higher up technical hires have good technical skills, which I think most developers would agree is a good idea.But their way of measuring this (the standard way) is bad, because it ignores the nuance a more experienced person has (seen so clearly here).On the gripping hand, I can&#x27;t help but feel the candidate did demonstrate a big failure to communicate, which is an important skill in itself.For example, listing SYN, SYN-ACK, ACK in hexadecimal is great for showing off, but is legitimately a bad answer to the question – as evidenced by the lack of understanding in the questioner. I also think some social graces might have got them further (e.g. &quot;Oh yeah, sure. Quicksort is O(N log N on average, and is generally a reasonable sort to chose, but I wanted to mention some other factors that are worth considering&quot;).At the end of the day:- it&#x27;s a hoop. Jump through it and get a fish, or don&#x27;t.- calibrate for your audience! This is a very important technical skill (and this test was unintentionally correct in its result IMHO).

Mediocre people can&#x27;t evaluate exceptional people. What mediocre people think is knowledgeable can&#x27;t be exceptional knowledge, by definition. If you want to find exceptional directors of engineering, you need an exceptional person worthy of being director of engineering to evaluate them.In seeking victory, not going beyond what everyone knows &#x2F;
 is not skilled.Victory in battle that all-under-heaven calls skilled &#x2F;
 is not skilled.Thus lifting an autumn hair does not mean &#x2F;
 great strength.Seeing the sun and moon does not mean a &#x2F;
 clear eye.Hearing thunder does not mean a keen ear.So-called skill is to be victorious over the easily &#x2F;
 defeated.Thus the battles of the skilled are without &#x2F;
 extraordinary victory, without reputation for &#x2F;
 wisdom and without merit for courage.- Sun Tzu Ch. 4 (Denma translation)

The Google person&#x27;s response to the answer in question 5 would have pretty much ended the call with &quot;can I talk to an actual engineer as you are doing poor keyword searches on my answers&quot;. I do admit to hearing the &quot;I will stab you&quot; voice in my head on that one[1]. This reflects really badly on Google.Plus the answer on question 8 is the best worded answer to using big-O improperly I have seen.1) <a href="http:&#x2F;&#x2F;www.urbandictionary.com&#x2F;define.php?term=I%20will%20stab%20you" rel="nofollow">http:&#x2F;&#x2F;www.urbandictionary.com&#x2F;define.php?term=I%20will%20st...</a>

What i think he means is counting the amount of set(1) bits in the array,

If you have a collection of data, and you want to know the number of 1 bits, and you want to do it with a minimum of resources... what is the process?For example, standard bit-shifting and masking the lowest bit to set a counter is one way to do this. Possibly there are other, faster ways, such as using a lookup table (a byte or more can be &quot;counted&quot; at a time). Of course, because so many people were doing this, intel added a popcnt instruction which probably is more efficient (faster) than either of the above, at the expense of more CPU real estate, heat, etc.Turns out counting 1 bits in a dataset is a super-important problem that shows up in a lot of situations.

&quot;how do you count the bits most efficiently?&quot;What does this even mean?

Its worst case is O(n^2) sure, so heap sort would be better if you know that your data would hit the worst case scenario of quick sort every time.

Just a nitpick:  There are versions of quicksort that perform well, O(n*log n), on already sorted data.  But there is always some worst case scenario where it can be O(n^2).

quicksort is O(nlogn) average case, and O(n2) on already sorted arrays.

Quicksort is O(n log n) average and O(n^2) worst case. Heapsort is O(n log n) for both average and worst case.

quicksort is O(n^2) which is definitely not the &quot;best big-O&quot; for sorting.

Absolutely not. The reason so many Google upper management types sport beards (case in point, Sundar Pichai), is the chronic lack of time from working nights as data-center technicians&#x2F;network engineers. How else can you explain Google&#x27;s up-time and low-latency? :)

&quot;Recruiter: wrong, it&#x27;s SYN, SYN-ACK and ACK; if Google is down you will need to know this to diagnose what the problem is. We will stop here because it&#x27;s obvious that you don&#x27;t have the necessary skills to write or review network applications. &quot;Because when Google is down, it&#x27;s typically upper management that fixes network issues. CEO level to directors of engineering are on pager duty most nights there.

Google's “Director of Engineering” Hiring Test