3 ms·
I think there is a lot of negative bias here, because people who do badly in algorithms and data structure questions, or dislike them for some other reason, are
by mc-lovin 13y ago
I think there is a lot of negative bias here, because people who do badly in algorithms and data structure questions, or dislike them for some other reason, are more likely to speak on the issue.
I think that these questions are good because they are difficult, and a smarter more capable person is much more likely to get them right. Notice that big tech companies have stopped with the "brainteaser" style questions like how many potholes in Manhatten, but there is no move away from algorithms and data structures questions. And I don't consider the format to be high pressure. Apart from the pressure inherent in having to answer the question on the spot, I never felt any pressure from the interviewer, they tended to be very polite and mild mannered. At worst they didn't understand or agree with my solution.
Telling people that these interview questions suck and they should look for jobs where they don't have to answer them, is really bad advice. Some of the best jobs (for most people) require them. If you face such an interview, the very best thing you can do is study. I probably did 100 interview questions in preparation for my job interviews.
- manish_gill 13y agoThe fact is that the companies failed to communicate when he explicitly asked for feedback. That has nothing to do with his performance in the interviews, unless these are people who simply don't reply because they thing the candidate is bad.
- dirktheman 13y agoInteresting point of view, which I don't share by the way. I think the right person for the job is not per definition the smartest or the best at solving equations during an interview. Finding the right person for the job is more about getting a good fit with the rest of the team. Being the smartest guy/girl on earth doesn't help you when you have a shitty personality or can't work together with the rest of the team. If I were to recruit a programmer, I'd look for past work (Github repos, online projects), accomplishments and personality, rather than test results that don't reflect the reality of the job in any way.
- mc-lovin 13y agoWell it's hard to quantify the relative importance of personality and cultural fit vs intelligence and programming ability. You can turn what you said around, and say that you can be the nicest, coolest person in the world, but if you are a fucking moron who can't code, you won't be very useful. But the big tech companies seem to have settled on the current process, and are therefore probably fairly happy with the people their interview process is giving them. As to looking at previous projects, that makes sense but it might be more time consuming to properly evaluate someone like that.
- randomdata 13y agoThe big tech companies haven't been happy with the people the interview process have been giving them – all we heard about for a long time is about how they couldn't find anyone to hire. Perhaps this settling you speak to marks the end of developer shortages in the industry?
- mc-lovin 13y agoNot really. These are two separate issues. If I am a gold panner, I might be very satisfied with my gold panning technique, and yet complain that there is a shortage of gold in the river.
- bane 13y agoI'm going to politely disagree on some points: Consider this scenario, your company will never need an algorithm designer. The jobs they need are to do basically glorified component integration. You have a candidate who is awesome at gluing together various libraries and other odds and ends into a working product. The things he's worked on are well known and well regarded, and he has 15 years out of college doing these things. He's never even had to think about the complexity of a red-black tree in all that time let alone implement one. He's solid, a great team lead, personable without any difficult to handle "quirks". He's been the core to the success of several products. In other words, he's the perfect match. So why filter him out by testing how much of an algorithm encyclopedia he is? It'd be like interviewing an algorithm guy by asking him how well he knows some web framework. At best it's non sequitur, at worst you won't get the employee you need. >Notice that big tech companies have stopped with the "brainteaser" style questions like how many potholes in Manhattan My understanding is that they did this due to a number of lawsuits. The courts felt that it presented a discriminating playfield, and most large companies have stopped this due to recommendations form their legal departments.
- abhiv 13y agoWhere can you find prep questions for technical interviews? Glassdoor or somewhere else?
- ccoggins 13y agoI'd recommend "Cracking the Coding Interview" by Gayle Laakmann McDowell. It's cheap at around 25 bucks on Amazon and has a pretty good cross section of questions. It isn't comprehensive by any means but I feel it's a good place to start. I think there is also a companion website that has more questions.