6 ms·

Is something wrong with software engineering interviews that one could reasonably need to spend months prepping for interviews? Leetcode problems can be fun and

by cebert 5y ago

Is something wrong with software engineering interviews that one could reasonably need to spend months prepping for interviews? Leetcode problems can be fun and interesting, but seldom represent problems and skills I employ daily as a more senior software engineer. Studying these problems won’t help me much at my current employer and likely wouldn’t after I passed the interview gauntlet at another firm.

nostrademons 5y ago
It's because of the time constraints required in an interview. At my job, the assumption is that you won't have enough context and experience to be effective in the job until you've been there for 6 months. You can't fit 6 months of prep into an interview. Good interview questions are ones that boil that 6mo+ of experience down to something that can be explained in 5 minutes and solved in 45. Ideally you'd just work for the company for a few months and then they have their signal, and that's how internships and contractor conversions work. Most people with experience are unwilling to quit a job for a 3-month trial period though.
- H8crilA 5y ago
  This. Also, 3 month trial periods instead of interviews are not going to happen for two reasons: a) there are still many people out there applying for jobs that couldn't code an append operation for a linked list, b) the cost that such trials would put on companies - if you ever mentored a junior you know the cost. It's worth it, but only because >90% of the recruits are valuable afterwards.
- heavyset_go 5y ago
  At-will employment means that any period is a trial period if the company wants it to be.
  - 0des 5y ago
  - rsanek 5y ago
    I've long thought that it might be interesting for an interview model to emerge where a trial period of weeks/months is the expectation. I think this might resolve much of the current frustration with programming inteviews. But I also recognize that doing so would be significantly different from how essentially all other companies operate, and if it weren't communicated clearly enough, could result in negative PR for a firm trying such an approach.
  - mandelbrotwurst 5y ago
    Yes, and if they pull that trigger too frequently they'll earn a reputation for firing people, which isn't a great way to attract talent.
- higeorge13 5y ago
  He interviewed as an engineering manager, his job will be probably people management, hiring and moving tickets around. Do engineering managers in meta do any technical work in order to require lc mindset? What 6 month experience do you expect from an new joiner engineering manager other that attending meetings and get to know the projects in high level and what do you expect to see in these interviews through lc questions?
anyfoo 5y ago
There are plenty of passionate software engineers who think those problems are actually fun. Independently of that (I don't think it's an immediate red flag if those problems are not fun for you, or not fun for you in that specific setting etc.), a lot of it is actually a glimpse into how working with that person might be. I.e. what ideas do they form, how do they communicate them, how do they respond to feedback, how do discussions about alternatives, problems etc. go, do they seek help when they need it and how... I don't think it's deniable that those are valuable qualities to test.
- fdgsdfogijq 5y ago
  Honestly I do interviews at FAANG, and all those qualities mean nothing. At the end of the day, if they solve the questions they will probably get an offer, and if they dont, they probably wont.
- alephnan 5y ago
  > There are plenty of passionate software engineers who think those problems are actually fun. There are plenty of passionate software engineers who don’t. Competitive programmers maybe. I like writing naive code that does something useful, where useful sometimes mean performance. When you revisit or just step away from the problem, you’ll have an aha moment of how to solve it more efficiently, or more elegantly. The Leetcode style interview expects finding that insight in the first 5-10 minutes of the interview. Here, more grinding increases the probability you arrive that rote insight. Rote memorization is not very fun, especially for software engineers who like assembling complex things from simple primitives and first principles.
  - anyfoo 5y ago
    So? I did write that not having fun with that is not a red flag for me, and listed qualities that can be assessed from such an interview independently of whether someone has fun with it or not. I do have fun with computer science problems, and it translates into an appetite to solving tricky technical problems in my job. I have peers that are similar, given that pre-COVID I remember us discussing other people's interview questions purely for the fun of discussing them. If the job specifically calls for tricky problem solving on a regular basis then, yes, I think a certain passion and appetite visible in the interview helps, but I'm well aware that that's not universal. For example, I assume (not that I know, feel free to correct me) that in frontend development of websites, something I can see is a lot of work and which I respect, other qualities are more important than a passion for intricate technical and/or computer science puzzles. (In fact, I think you are right that there certainly are subfields where trying to come up with an intricate solution, instead of a known and proven one, can be rather detrimental.)
actually_a_dog 5y ago
Yes, there is something wrong and you just said it in your last sentence.
shrimp_emoji 5y ago
(Chronic) Licensing and accreditation are caching the result of an operation. Since no such licensing or accreditation exists in software, the operation has to be repeated over and over, which is difficult.
- Something1234 5y ago
  And most licensing and accreditation is a joke. It's never firm enough. These certs can be gamed. Colleges are grade inflated and too many idiots are passed.
gleenn 5y ago
The very last bit of the article interestingly stated that they author wouldn't spend time doing prep again, they would only make sure they applied with a referral. I definitely agree, it's kind of a blight on our industry that the interview process is so contorted compared to the actual work we're supposed to do.
- vln 5y ago
  I think it's because he's on the eng management track. I don't think they're asking hard leetcode questions.
  - R0b0t1 5y ago
    Then why is he making ~$500k/yr?
- neilv 5y ago
  I thought maybe the author meant that they'd retain enough of the months of prep that they'd already done.
  - lumost 5y ago
    Generally, once you've practiced enough to clear most tech rounds once - it's not too difficult to do it again. If I'm going in for an interview then I might still need to spend ~1-2 weeks practicing again to do well. Earlier in my career I spent months practicing LC.
    - scruple 5y ago
      Eh, I'm 8 years into the same company now, have been shipping the entire time across multiple teams, wearing multiple hats, etc., and am 15 years into my career. I'd have to put some non-trivial effort into refreshing my memory of a great deal of the sorts of things that are (anecdotally) commonplace in interview settings today. And with a full-time job and 3 young kids, time is hard to come by. 8 years ago, I spent time with CtcI, etc., and was immensely prepared to interview but, for me, that's all largely lost to the ages now.
jmchuster 5y ago
Maybe it's because i know a lot of doctor friends, but months of prepping to then get paid more than they do seems way unfair in my favor.
- sarchertech 5y ago
  The median software engineer salary isn't anywhere near the median doctor's salary. I don't know that most engineers could make it through FAANG interviews even with months of prep.
  - jmchuster 5y ago
    I was going off more of FAANG salaries, comparing the work you put in to get a FAANG job vs the work you put in to get a doctor job. Huh, i guess it probably is true that "most" engineers couldn't make it, but i guess my viewpoint is skewed because "most" of my engineer friends could and do make it, with "months" of prep being an upper bound, more closer to "weeks". So, it is a bit selective, but i don't know how selective it is relative to people who are able to get into medical school.

by cebert 5y ago

nostrademons 5y ago

It's because of the time constraints required in an interview. At my job, the assumption is that you won't have enough context and experience to be effective in the job until you've been there for 6 months. You can't fit 6 months of prep into an interview. Good interview questions are ones that boil that 6mo+ of experience down to something that can be explained in 5 minutes and solved in 45. Ideally you'd just work for the company for a few months and then they have their signal, and that's how internships and contractor conversions work. Most people with experience are unwilling to quit a job for a 3-month trial period though.

H8crilA 5y ago
This. Also, 3 month trial periods instead of interviews are not going to happen for two reasons: a) there are still many people out there applying for jobs that couldn't code an append operation for a linked list, b) the cost that such trials would put on companies - if you ever mentored a junior you know the cost. It's worth it, but only because >90% of the recruits are valuable afterwards.
heavyset_go 5y ago
At-will employment means that any period is a trial period if the company wants it to be.
- 0des 5y ago
- rsanek 5y ago
  I've long thought that it might be interesting for an interview model to emerge where a trial period of weeks/months is the expectation. I think this might resolve much of the current frustration with programming inteviews. But I also recognize that doing so would be significantly different from how essentially all other companies operate, and if it weren't communicated clearly enough, could result in negative PR for a firm trying such an approach.
- mandelbrotwurst 5y ago
  Yes, and if they pull that trigger too frequently they'll earn a reputation for firing people, which isn't a great way to attract talent.
higeorge13 5y ago
He interviewed as an engineering manager, his job will be probably people management, hiring and moving tickets around. Do engineering managers in meta do any technical work in order to require lc mindset? What 6 month experience do you expect from an new joiner engineering manager other that attending meetings and get to know the projects in high level and what do you expect to see in these interviews through lc questions?

anyfoo 5y ago

There are plenty of passionate software engineers who think those problems are actually fun. Independently of that (I don't think it's an immediate red flag if those problems are not fun for you, or not fun for you in that specific setting etc.), a lot of it is actually a glimpse into how working with that person might be. I.e. what ideas do they form, how do they communicate them, how do they respond to feedback, how do discussions about alternatives, problems etc. go, do they seek help when they need it and how... I don't think it's deniable that those are valuable qualities to test.

fdgsdfogijq 5y ago
Honestly I do interviews at FAANG, and all those qualities mean nothing. At the end of the day, if they solve the questions they will probably get an offer, and if they dont, they probably wont.
alephnan 5y ago
> There are plenty of passionate software engineers who think those problems are actually fun. There are plenty of passionate software engineers who don’t. Competitive programmers maybe. I like writing naive code that does something useful, where useful sometimes mean performance. When you revisit or just step away from the problem, you’ll have an aha moment of how to solve it more efficiently, or more elegantly. The Leetcode style interview expects finding that insight in the first 5-10 minutes of the interview. Here, more grinding increases the probability you arrive that rote insight. Rote memorization is not very fun, especially for software engineers who like assembling complex things from simple primitives and first principles.
- anyfoo 5y ago
  So? I did write that not having fun with that is not a red flag for me, and listed qualities that can be assessed from such an interview independently of whether someone has fun with it or not. I do have fun with computer science problems, and it translates into an appetite to solving tricky technical problems in my job. I have peers that are similar, given that pre-COVID I remember us discussing other people's interview questions purely for the fun of discussing them. If the job specifically calls for tricky problem solving on a regular basis then, yes, I think a certain passion and appetite visible in the interview helps, but I'm well aware that that's not universal. For example, I assume (not that I know, feel free to correct me) that in frontend development of websites, something I can see is a lot of work and which I respect, other qualities are more important than a passion for intricate technical and/or computer science puzzles. (In fact, I think you are right that there certainly are subfields where trying to come up with an intricate solution, instead of a known and proven one, can be rather detrimental.)

actually_a_dog 5y ago

Yes, there is something wrong and you just said it in your last sentence.

shrimp_emoji 5y ago

(Chronic) Licensing and accreditation are caching the result of an operation. Since no such licensing or accreditation exists in software, the operation has to be repeated over and over, which is difficult.

Something1234 5y ago
And most licensing and accreditation is a joke. It's never firm enough. These certs can be gamed. Colleges are grade inflated and too many idiots are passed.

gleenn 5y ago

The very last bit of the article interestingly stated that they author wouldn't spend time doing prep again, they would only make sure they applied with a referral. I definitely agree, it's kind of a blight on our industry that the interview process is so contorted compared to the actual work we're supposed to do.

vln 5y ago
I think it's because he's on the eng management track. I don't think they're asking hard leetcode questions.
- R0b0t1 5y ago
  Then why is he making ~$500k/yr?
neilv 5y ago
I thought maybe the author meant that they'd retain enough of the months of prep that they'd already done.
- lumost 5y ago
  Generally, once you've practiced enough to clear most tech rounds once - it's not too difficult to do it again. If I'm going in for an interview then I might still need to spend ~1-2 weeks practicing again to do well. Earlier in my career I spent months practicing LC.
  - scruple 5y ago
    Eh, I'm 8 years into the same company now, have been shipping the entire time across multiple teams, wearing multiple hats, etc., and am 15 years into my career. I'd have to put some non-trivial effort into refreshing my memory of a great deal of the sorts of things that are (anecdotally) commonplace in interview settings today. And with a full-time job and 3 young kids, time is hard to come by. 8 years ago, I spent time with CtcI, etc., and was immensely prepared to interview but, for me, that's all largely lost to the ages now.

jmchuster 5y ago

Maybe it's because i know a lot of doctor friends, but months of prepping to then get paid more than they do seems way unfair in my favor.

sarchertech 5y ago
The median software engineer salary isn't anywhere near the median doctor's salary. I don't know that most engineers could make it through FAANG interviews even with months of prep.
- jmchuster 5y ago
  I was going off more of FAANG salaries, comparing the work you put in to get a FAANG job vs the work you put in to get a doctor job. Huh, i guess it probably is true that "most" engineers couldn't make it, but i guess my viewpoint is skewed because "most" of my engineer friends could and do make it, with "months" of prep being an upper bound, more closer to "weeks". So, it is a bit selective, but i don't know how selective it is relative to people who are able to get into medical school.