|
TooMuchAbstraction posted:Something else that may be working against you is simply that I'd guess 90+% of interviewees get rejected out of hand. I know at Google the policy is that if even one interviewer doesn't like you, then that's a no (but please try again in 6 months or a year). And interviewers will tend to err on the side of saying no if they're on the fence, because that means maintaining the status quo and that feels safe. So basically you have to wow absolutely everyone in order to pass. That's not the policy at Google, at least not for the last 4 years I've been interviewing, you can get hired with an interviewer giving a strong negative -- but it's pretty uncommon to get a strong negative while everyone else is glowing.
|
# ? Jun 28, 2018 20:23 |
|
|
# ? May 10, 2024 02:02 |
|
Volguus posted:So this is exhibit A of why the interviewing process is so hosed up in software. "You have to have experience with technology X, version 1.2.3 otherwise you're not even invited to the interview". And you, Keetron, are not alone in this approach. Almost everyone does the same thing. You can also get into consulting. If you get the right kind of gig, you can bounce around from client to client and learn new stuff on their dime, while pretending to already be an expert at it.
|
# ? Jun 28, 2018 20:24 |
|
New Yorp New Yorp posted:You can also get into consulting. If you get the right kind of gig, you can bounce around from client to client and learn new stuff on their dime, while pretending to already be an expert at it. Yeah, that's the best of both worlds: make the franekestein application, run away before it explodes, list poo poo on your resume, all on company time/dime.
|
# ? Jun 28, 2018 20:27 |
|
apseudonym posted:That's not the policy at Google, at least not for the last 4 years I've been interviewing, you can get hired with an interviewer giving a strong negative -- but it's pretty uncommon to get a strong negative while everyone else is glowing. Hm, maybe it depends on the department or something then.
|
# ? Jun 28, 2018 20:33 |
|
apseudonym posted:Can you roughly reproduce your code here? I know there are things I can change and some of the method calls are likely off (I've been working with Scala lately but still don't feel strong enough to choose it over Java) but this is roughly what I produced during the interview: code:
|
# ? Jun 28, 2018 20:44 |
|
Good Will Hrunting posted:I know there are things I can change and some of the method calls are likely off (I've been working with Scala lately but still don't feel strong enough to choose it over Java) but this is roughly what I produced during the interview: This doesn't seem to match the problem you described, fwiw. I'm not sure if "if(found[entry.value()] = true)" is a translation error(that part of your code doesn't make sense), but you shouldnt need a final loop as you can return early.
|
# ? Jun 28, 2018 20:59 |
|
TooMuchAbstraction posted:Hm, maybe it depends on the department or something then. He's right though. If one person is extremely negative while everyone else is positive the hiring committee usually tries to dig into the feedback and figure out why, including how much the interviewer deviated from peers in past interviews. HC ultimately has the biggest influence by far, there's no "one person didn't like you automatic fail" rule.
|
# ? Jun 28, 2018 21:00 |
|
apseudonym posted:This doesn't seem to match the problem you described, fwiw. What? Sure it does. Maybe I explained it poorly but let's use "banana" as an example: First for loop counts # of appearances per character in the string: (b -> 1, a -> 3, n -> 2) Second for loop flags an array using the array index as the value to store so my array found after that loop looks like [true, true, true]. The found[entry.getValue() == true] is checking whether this specific count has already been flagged, which denotes a failure (because we can only have 1 of each unique count). While I was writing code I thought we'd need that third for loop to look for gaps in that constraint. You are correct in that the last loop is not necessary. If you see an index that is greater than the map's size, we can terminate. I couldn't say that with certainty as I was wrapping up my solution but realized afterwards.
|
# ? Jun 28, 2018 21:09 |
|
Good Will Hrunting posted:I know there are things I can change and some of the method calls are likely off (I've been working with Scala lately but still don't feel strong enough to choose it over Java) but this is roughly what I produced during the interview: I don't think this line works. code:
code:
|
# ? Jun 28, 2018 21:11 |
|
mrmcd posted:He's right though. If one person is extremely negative while everyone else is positive the hiring committee usually tries to dig into the feedback and figure out why, including how much the interviewer deviated from peers in past interviews. I read this on memegen today so it is probably true
|
# ? Jun 28, 2018 21:12 |
|
Jose Valasquez posted:If the question was about exactly 1 of any letter, exactly 2 of any letter, exactly 3 of any letter, etc. then I think this is fine? That wasn't how I understood the problem at first but upon rereading it maybe that's what it was. Yes, that was the problem. Also yeah, I'm sure there were other bugs too. This was the third problem I was tackling in the span of just over 50 minutes in a notepad with no syntax highlighting or anything, my brain was absolutely fried. I didn't have much time to review what I had written at all, or like, actually think about it besides feeling like I was racing a clock.
|
# ? Jun 28, 2018 21:13 |
|
Good Will Hrunting posted:Yes, that was the problem. If I was the interviewer I wouldn't take off much credit for the ++ thing personally, but I'm also not an interviewer so
|
# ? Jun 28, 2018 21:15 |
|
Good Will Hrunting posted:What? Sure it does. Maybe I explained it poorly but let's use "banana" as an example: Alright, that makes more sense. As a nit please don't do == true and == false, it's minor but looks super amateurish. Your array size is incorrect either way, consider "aaaaaa", if you use the size of the map you'll try and access found[6], which is out of bounds. You'd have to use the max count. E: and other things people pointed out
|
# ? Jun 28, 2018 21:21 |
|
So basically I am expected to come up with a flawless solution with zero bugs that I have the time to go through and test in the constraints of 30 minutes? I mean sure, it's not a particularly hard problem, but that still seems unrealistic to me and once again beyond the whole "we want to see how you solve a problem" that people said these interviews were about. If I had time, I would have stepped through and seen most of the mistakes I made pretty quickly. But I didn't have time. So am I supposed to code... faster?
|
# ? Jun 28, 2018 21:28 |
|
My new workplace in Chicago is looking for a full stack senior developer with experience in things like rails, react, and/or elixir. I don't want to post too many identifiable details, but if you're interested PM me and I can send you the link to the job description and tell you more about it.
|
# ? Jun 28, 2018 21:31 |
|
Good Will Hrunting posted:So basically I am expected to come up with a flawless solution with zero bugs that I have the time to go through and test in the constraints of 30 minutes? I mean sure, it's not a particularly hard problem, but that still seems unrealistic to me and once again beyond the whole "we want to see how you solve a problem" that people said these interviews were about. If I had time, I would have stepped through and seen most of the mistakes I made pretty quickly. But I didn't have time. So am I supposed to code... faster? It doesn't have to be flawless mistakes are kinda expected, but if you detect them without me asking it's a big plus, if I have to prod you with leading questions for twenty minutes it's a minus. Wanting to see how you solve a problem is both how you come up with a solution as well as how you identify bugs in your approach or code.
|
# ? Jun 28, 2018 21:32 |
|
This is an old tweet, but maybe it will help you feel better https://twitter.com/mxcl/status/608682016205344768?lang=en edit: have some more
|
# ? Jun 28, 2018 21:33 |
|
apseudonym posted:Wanting to see how you solve a problem is both how you come up with a solution as well as how you identify bugs in your approach or code. I would have stepped through multiple example cases, starting with a correct one, then like you said something ridiculous with all the same characters, then one with something like 1 char and 3 chars to make sure I got steps missed in the sequence properly returning false. But unfortunately I had to so some stupid palindrome poo poo and a binary tree question first.
|
# ? Jun 28, 2018 21:36 |
|
apseudonym posted:That's not the policy at Google, at least not for the last 4 years I've been interviewing, you can get hired with an interviewer giving a strong negative -- but it's pretty uncommon to get a strong negative while everyone else is glowing. I got one meh and everyone else was chuffed, took like 4 weeks to get the meh answer and the "job had gone" by the time the response was filed. That's with Google NYC.
|
# ? Jun 28, 2018 21:38 |
|
This makes me feel marginally less useless after my first tech screen round+ interviews have been rejections, thanks! Still wondering if I chose the wrong career path and that imposter syndrome is strong as hell right now.
|
# ? Jun 28, 2018 21:45 |
|
The Fool posted:This is an old tweet, but maybe it will help you feel better Qualified people get rejected all the time and nobody should take it as a judgment of their ability or self worth. That being said the Homebrew guy admitted that he isn't very good at designing software and that he is an rear end in a top hat. Given those two things the hiring process worked perfectly in this scenario. Also, the number of people at Google using Macbooks is way lower than 90% and those that do generally aren't using homebrew (at least not for work)
|
# ? Jun 28, 2018 21:45 |
|
MrMoo posted:I got one meh and everyone else was chuffed, took like 4 weeks to get the meh answer and the "job had gone" by the time the response was filed. That's with Google NYC. If it was for a specific team, as opposed to general SWE where team matching comes after, then it's more on the highering manager, and it sounds like they found someone else Good Will Hrunting posted:I would have stepped through multiple example cases, starting with a correct one, then like you said something ridiculous with all the same characters, then one with something like 1 char and 3 chars to make sure I got steps missed in the sequence properly returning false. But unfortunately I had to so some stupid palindrome poo poo and a binary tree question first. Yeah that sucks sometimes interviewers don't do good at managing the time or they expected you to be faster than you were.
|
# ? Jun 28, 2018 21:46 |
|
Good Will Hrunting posted:So basically I am expected to come up with a flawless solution with zero bugs that I have the time to go through and test in the constraints of 30 minutes?
|
# ? Jun 28, 2018 22:13 |
|
Jose Valasquez posted:That being said the Homebrew guy admitted that he isn't very good at designing software and that he is an rear end in a top hat. Given those two things the hiring process worked perfectly in this scenario. Also...homebrew is nice and all, but it's not like genius level software? It's a package management system. Used by lots of people != the guy who wrote it is a 100% certified genius hire. Like, how many people use that Javascript leftpad module? Also also, from the wiki for homebrew - 'Homebrew does not honor the default privileges of /usr/local; directory ownership is changed from root with group permissions for the wheel group to the installing user and the "admin" group' Ummmm. Maybe it's my Linux/commercial Unix background but that's making me dry heave. feedmegin fucked around with this message at 23:21 on Jun 28, 2018 |
# ? Jun 28, 2018 23:18 |
|
feedmegin posted:Also...homebrew is nice and all, but it's not like genius level software? It's a package management system. Making a working package manager seems to be shockingly difficult. Anything remotely resembling a success should be lauded. (Lauded doesn’t necessarily mean hired though, I agree.)
|
# ? Jun 28, 2018 23:24 |
|
apseudonym posted:It doesn't have to be flawless mistakes are kinda expected, but if you detect them without me asking it's a big plus, if I have to prod you with leading questions for twenty minutes it's a minus. I give questions to SREs like "You get paged that Service X is delivering a 500 Internal Server error, what do you do check first?", and while "Check the logs" is an ok answer, an even better one is "In my experience the most common reason for internal service errors is resource exhaustion, misconfiguration, and buggy code, so I would check for all those things in that order. To check for resource exhaustion I would ...". This demonstrates a systematic approach, which matters more to me than someone who knows a lot but flounders around picking things to try.
|
# ? Jun 28, 2018 23:33 |
|
Maybe I should apply for Junior jobs since my 4.5 years of experience is Not Good experience for more than half of my tenure?
|
# ? Jun 28, 2018 23:36 |
|
Good Will Hrunting posted:Maybe I should apply for Junior jobs since my 4.5 years of experience is Not Good experience for more than half of my tenure? Nah you just need to chill and keep doing interviews. Every place is really heavily weighted for false negatives. I got like 3 or 4 rejections and 2 offers before the Google offer.
|
# ? Jun 29, 2018 00:20 |
|
minato posted:This. The solution doesn't give as much signal as the process. Expressing the thought process reveals how you deal with stuff you've never seen before; throw spaghetti at the wall or systematically break down the problem. Am I crazy for thinking that is a wrong answer? Given only the facts of 500s I would definitely first go digging for more info before making any assumptions about the root cause. Like maybe after working on a product for several months you could have developed a sane heuristic of “500s => very likely resource exhaustion” but it would be only for that service and the wrong approach without the experience to back it up.
|
# ? Jun 29, 2018 07:08 |
|
Eggnogium posted:Am I crazy for thinking that is a wrong answer? Given only the facts of 500s I would definitely first go digging for more info before making any assumptions about the root cause. Like maybe after working on a product for several months you could have developed a sane heuristic of “500s => very likely resource exhaustion” but it would be only for that service and the wrong approach without the experience to back it up. Also put an item on the techdebt backlog to implement proper error codes.
|
# ? Jun 29, 2018 08:00 |
|
Eggnogium posted:Am I crazy for thinking that is a wrong answer? Given only the facts of 500s I would definitely first go digging for more info before making any assumptions about the root cause. Like maybe after working on a product for several months you could have developed a sane heuristic of “500s => very likely resource exhaustion” but it would be only for that service and the wrong approach without the experience to back it up. Minato's point is it's more nuanced than just a solution and lays out a plan to find the root of the problem. Interview answers don't *need* to be technically correct as long as you explain your thought process and back it up sufficiently.
|
# ? Jun 29, 2018 12:17 |
|
Good Will Hrunting posted:Yep, it's Bridgewater. My roommie told his boss before he started interviewing and his boss was cool with him taking a day off per week to interview for over a month. Place may be culty, but that's a cool as gently caress culture. Of course he had been there 3 years and they had established a relationship of trust, and he was accomplishing his work still and agreed to stay an extra bit to train a potential new hire or offload his work but that's still pretty awesome, imo. i've known people who work at both amazon and bridgewater, and tbh it's kind of like comparing the mormons to the heaven's gate cult - amazon may dress you up in weird underwear and push you to succeed, but bridgewater will cut your balls off and make you drink poison. Bruegels Fuckbooks fucked around with this message at 13:49 on Jun 29, 2018 |
# ? Jun 29, 2018 13:02 |
|
Keetron posted:Also put an item on the techdebt backlog to implement proper error codes. That is the proper error code to send to clients when something falls over while serving an HTTP request, though? Doh004 posted:Minato's point is it's more nuanced than just a solution and lays out a plan to find the root of the problem. Interview answers don't *need* to be technically correct as long as you explain your thought process and back it up sufficiently. I think that when I approach a novel troubleshooting problem with preconceived notions of what probably went wrong that I often end up wasting time disproving my guesswork before making real progress on the problem, but that could be faulty/biased memory - certainly haven't been rigorously keeping track.
|
# ? Jun 29, 2018 14:21 |
|
Munkeymon posted:I think that when I approach a novel troubleshooting problem with preconceived notions of what probably went wrong that I often end up wasting time disproving my guesswork before making real progress on the problem, but that could be faulty/biased memory - certainly haven't been rigorously keeping track. I think that's a really important thing to consider (not letting your biases force you into tunnel vision) - but I expect an experienced engineer to use their previous experiences to inform their thought processes moving forward. That said, I also expect people to begin answering those types of questions by gathering more information and requirements up front (it'll almost never hurt to ask more questions).
|
# ? Jun 29, 2018 14:48 |
|
mrmcd posted:Nah you just need to chill and keep doing interviews. Every place is really heavily weighted for false negatives. I got like 3 or 4 rejections and 2 offers before the Google offer. Chill is the operative word here. I can't believe I stumbled on a simple max stock prices array question yesterday. I'm really starting to stress/burn myself out. Broadening my horizons outside of mostly start-ups might be a good idea. I got some Goon recommendations via PM that I've added to my list but after the 4-5 still in the pipeline that I'm working on now, I'll probably take a brief break from all of this to decompress. It really is almost a second full-time job.
|
# ? Jun 29, 2018 15:19 |
|
Just want to second what someone else said earlier in the thread: interviews are not about you or your intrinsic worth, they're about what you and a few employees of a business think of a possible employment relationship. An offer would feel validating and pleasant (and the structure of many of these interviews is essentially akin to a pop quiz or entrance exam, which can kind of throw at least me into a "school days, validate me for being smart plz" mindset) but no offer shouldn't be taken as a much more ambiguous signal.
|
# ? Jun 29, 2018 15:49 |
|
... is it weird to say that I feel like the next recession is coming on, that the "animal spirits" have changed? I get this feeling like the next time I'm searching for a job, there will be a bazillion formerly-employed and what's more formerly-quite-employable people very actively searching for jobs and the filters and processes will probably be even worse as a result. if I were william blake I'd hallucinate a vision and then write a poem about beholding a chilling figure holding a "will code for food" sign
|
# ? Jun 29, 2018 15:55 |
|
That's a terrifying thought and does wonders for my already increasing stress Anyway, here's a loosely reproduced version of my solution to "Given an array of integers and a target number, return a set of pairs of integers in the array that add to target" (I'm trying these again in a similar condition - in a notepad with no syntax or highlighting) E: I asked and yes, numbers can be duplicated E2: When I was first solving the problem I assumed it was Pairs of indices (not sure why I thought this, probably because it makes the problem more complicated given the clarity of the duplicated number constraint and I always assume they're trying to trick me with a "gotcha") so I began with a mapping to preserve the locations and as I was explaining this and had already coded all this poo poo on the whiteboard it would have been hard to have erase it. code:
Good Will Hrunting fucked around with this message at 16:42 on Jun 29, 2018 |
# ? Jun 29, 2018 16:30 |
|
prisoner of waffles posted:... is it weird to say that I feel like the next recession is coming on, that the "animal spirits" have changed? No, it's not weird and in fact supported by indicators like the yield curve. Makes me want to job hunt even though I haven't been here that long because I don't wanna be stuck here through the next recession. Not that it's bad or anything - just that I know I can do better. Should probably also sell my place, the thought of which makes interviewing seem downright pleasant.
|
# ? Jun 29, 2018 16:32 |
|
|
# ? May 10, 2024 02:02 |
|
May I introduce you all to a little thing I like to call democratic socialism?
|
# ? Jun 29, 2018 16:37 |