Algorithms interviews: theory vs. practice https://danluu.com/algorithms-interviews/ …
-
-
Replying to @danluu
Thoughtful post as always, but I don’t agree with the premise that “algorithms questions are important to big companies because *fancy algorithms* are important there”.
2 replies 0 retweets 0 likes -
Replying to @josephkaptur @danluu
It’s pretty easy to come up with “not bad” algorithms questions, which lets question banks be large, which means memorization matters less.
1 reply 0 retweets 0 likes -
Replying to @josephkaptur
I think algorithms filters can't be that good or Centaur couldn't have had the most productive engineers I've worked with while not having an algorithms filter at all. I would agree that algorithms questions can attempt to filter for useful qualities, but
1 reply 0 retweets 0 likes -
Replying to @danluu @josephkaptur
Dan Luu Retweeted Steve Canon
IMO they have a very high false positive and false negative rate and I think that there are more important things to filter for, e.g., what
@stephentyrone refers to herehttps://twitter.com/stephentyrone/status/1214194801496727553 …Dan Luu added,
Steve Canon @stephentyroneReplying to @stilescrisis @danluuThe cost of a bad hire (in the sense of algorithmic boneheadedness) is time spent by senior people on code reviews bringing them up to speed. The cost of a bad hire (in the toxic/abusive personality sense) is the complete destruction of the team.2 replies 0 retweets 1 like -
Replying to @danluu @stephentyrone
I definitely didn’t mean to imply that algorithms questions are “good”, period. I just wanted to give advantages other than “we need people to actually write fancy algorithms”.
1 reply 0 retweets 0 likes -
One of the advantages is that you can come up with a lot of different questions that are recognizably “algorithms questions”, which makes it hard to memorize them all. I think “have your friends told you our questions, and have you memorized the answers” is a bad criterion.
1 reply 0 retweets 0 likes -
I completely agree with that there are a ton of disadvantages/problems/biases, and, at best, should only be a component of a hiring process.
1 reply 0 retweets 0 likes -
Replying to @josephkaptur @stephentyrone
Again just IMO, but I think memorization isn't a huge problem for most non-FB/Google companies? I'm told Pivotal/CF has a single standard question for all phone screens (onsite is pair programming with someone on "actual" work), seems to work pretty well for them.
2 replies 1 retweet 3 likes -
I'm giving them as an example because they're big, so it's probably not super hard for people to find the question if they want. For onsites, Heap, Hunch (and later part of eBay NYC) had a single question, seemed to work pretty well (maybe also Matasano? Not sure about them).
1 reply 0 retweets 0 likes
I don't think FB or Google could get away with that, but I think companies up to the size of Pivotal/CF includes most companies. I don't disagree that there are downsides to having a single standard question (people can and will cheat), but in practice it seems to have worked?
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.