I love this idea to check how quickly a candidate can learn: if blocked on some coding/refactoring question, show the candidate how to do it, erase everything and ask him to redo it alone.
Stop Interviewing with Leet Code
261–270 of 675 posts
Re: Stop Interviewing with Leet Code
#262A lot of comments theorize about what characteristic - for the most part other than coding skill - these kinds of interviews select for. Mostly it's to justify their use. OK, then: is there any evidence that this style of interview effectively or efficiently selects for any particular characteristic that matters? If these companies were really as data driven as they all claim to be, they'd rigorously analyze predictors of success (whatever that means) within the system they've built, and then design an interview process that can rationally be expected to select for those characteristics. Having worked at one, interviewed at another, and heard lots of stories about the rest, I don't get the impression that any of them have actually done that.
Re: Stop Interviewing with Leet Code
#263All of these supposed "flaws" of leetcode are by design. Big companies want people who are smart enough to do the work, but obedient enough to put up with all the bullshit that comes with working at a big company. What person better matches that than someone who's able and willing to study for and pass a tech version of the SAT? Every anti-leetcode article I read is some version of "leetcode is bad because it measure…
It'd be a fun social experiment if one of the big tech companies replaced the leetcode-style rounds with something arbitrary. Let's say: The "jumping jacks" round. You have to turn your webcam on and do 100 jumping jacks in 60 seconds. Only then will you potentially advance to the next round. The person watching on the other end can evaluate how far over 100 you got, whether or not your form matches best practice, an…
Pretty awesome pace if you can do a 100 in a minute.
Re: Stop Interviewing with Leet Code
#264Earlier quoted context omitted.
Every unicorny startup I've interviewed w/ that had a standard big-tech interview loop (1 LC style screener + final w/ 2-3 rounds of LC + 1 system design) was chock full of ex-big tech engineers and managers, replete with stories about wanting a faster paced / dynamic environment. So yes, that absolutely is who they are recruiting. It probably comes down to no more than believing big tech companies have the very best…
I wonder which came first? Did they hire those people because they implemented leetcode style interviews, or did they implement those interviews because they originally hired or were founded by ex-bigco folks that simply did what they knew?
I notice quite a few early stage startups that recruit w/ large initial funding rounds (ie. a $20-40M series A) almost invariably have a note about how their founders are ex-FAANG. I can't help but believe it helps a ton w/ funding, professional networks, team building, etc.
Again, no different than having an Ivy League education. There are big advantages to scope/reach/opportunities in the industry.
Re: Stop Interviewing with Leet Code
#265Earlier quoted context omitted.
It seems like a huge inefficiency in the economy that thousands of people are studying and practicing for an entrance exam that has no value in the real work that the companies actually do. But overall that may be a good thing for society. If Google fired all their leet coders and replaced them with real engineers they would not need as many. And the biggest problem may be that there is a limited number of real engin…
> replaced them with real engineers nice gatekeeping. Shame on them for working to get a job they want.
I'm not saying that they should be ashamed for being leet coders. I'm just pointing out the inefficiency of the system. It isn't unique to software. Historically, lot of jobs have required irrelevant accomplishments or certifications before employment.
Re: Stop Interviewing with Leet Code
#266Earlier quoted context omitted.
It seems like a huge inefficiency in the economy that thousands of people are studying and practicing for an entrance exam that has no value in the real work that the companies actually do. But overall that may be a good thing for society. If Google fired all their leet coders and replaced them with real engineers they would not need as many. And the biggest problem may be that there is a limited number of real engin…
In your opinion, what's the difference between a leetcoder and a real engineer? In my mind, a real engineer really shines in the non technical aspect of things, like coordination, communication, prioritization, and getting hard questions answered. But that's just me, I'm curious what everyone else's experiences are.
All of that non-technical stuff wont help you when the bridge you signed off on collapses because you had no idea what the blueprint was actually saying.
I think a bunch of this "engineers don't actually need to know stuff" comes from a generation that never had to deal with things that might kill people if it fails.
Re: Stop Interviewing with Leet Code
#267All of these supposed "flaws" of leetcode are by design. Big companies want people who are smart enough to do the work, but obedient enough to put up with all the bullshit that comes with working at a big company. What person better matches that than someone who's able and willing to study for and pass a tech version of the SAT? Every anti-leetcode article I read is some version of "leetcode is bad because it measure…
It'd be a fun social experiment if one of the big tech companies replaced the leetcode-style rounds with something arbitrary. Let's say: The "jumping jacks" round. You have to turn your webcam on and do 100 jumping jacks in 60 seconds. Only then will you potentially advance to the next round. The person watching on the other end can evaluate how far over 100 you got, whether or not your form matches best practice, an…
I know it’s a jokey example but it has insight.
First, it’s probably illegal in the USA due to ADA unless you could show that it tested something physically related to the job. Example: a big company is required to make accommodation for a, say, qualified analyst who is blind, but can reject a blind candidate for a job that required driving.
But after 45+ years of hiring programmers, the world hasn’t yet figured out which factors are germaine and which are not.
Re: Stop Interviewing with Leet Code
#268Last xmas I worked on a side project that ended up requiring a moderately substantial algorithm; much more involved than anything I've ever had to do in a LeetCode interview - there's absolutely no way i'd be able to produce it under those circumstances. Yet, the algorithm wasn't the hard part. The hard part was all the exploration I had to do around the problem to get to the point where I understood the constraints…
I'm curious - could you tell us what the algorithm was?
It was somewhere between a sudoku solver and wave function collapse. Had to work out what tiles to use to accurately reproduce the slopes for rendering a Mario Maker 2 level. Most of the level tiles are just X/Y/TileID, but for slopes all you know is the start/end nodes. So you have to look at all the adjacent tiles to work out the correct TileID to use for each X/Y paid within the bounding box of the slope. But there's also ambiguity where there are multiple possible tiles with the same edge properties, so I had to work out the right priority/weighting to favour certain tiles over others.
I had to derive and then write out all the edges rules by hand, which was many hundreds of lines.
Re: Stop Interviewing with Leet Code
#269The best interview we've used was sharing a simple but very not idiomatic Python file with working (but slow) code and tests at the bottom. The task was to refactor and speed it up. This allows seeing the actual thought process and some basic skills, while at the same time being something any good dev could do in 20min without much pressure.
I tend toward this, especially with some canned performance report. Sometimes I take something relevant from their GitHub and introduce a glaring bug, like an impossible dependency or infinite loop.
Re: Stop Interviewing with Leet Code
#270> All starts with showing some code, a class that does some stuff and its corresponding tests. The code is not glaringly bad, but it’s also not great on purpose. [And all that follows]
This is worse than LeetCode in my opinion. Because all it is, is a shallow copy of LeetCode. You've constructed a puzzle by laying out a picture and cutting out particular pieces. It's "find the differences" between what you've given them and the image in your mind.
> If I’m hiring a landscaper, I’m not gonna ask them to tell me about the classification of ficus in Fiji, or the specific reproduction period of Douglas Fir in the West Coast. I’m gonna ask them to trim a tree and see if the result suits me.
And the landscaper will walk. They will not work for free. They'll give you references, they'll show you pictures of work they've done before, but they won't do work for you.