It's just a baseline, to filter out applicants who can't do the "secret handshake". Ppl will still need to be interviewed and can be trained afterwards
"So how'd you make the building even taller? Thought you had limited resources" "Oh we just took the stuff from below and put it at top" "You did what-"
I ask them to talk about something they built in the past. I get into the details, challenges, and choices. It becomes obvious pretty quickly if they understand what they did, rather than just put some keyword on a resume.
I do the same. This “leetcode” nonsense has to stop. I’d rather see the person’s eyes light up talking about their project and how engaged they are when we “what if” the use case or their assumptions. Companies give 6 round leetcode interviews for their mundane “make a corporate dashboard” level jobs. Absolutely counterproductive.
Same, that's exactly how I hire all my devs. I also volunteer at 2 code camps doing mock interviews, and it is super obvious who the best coders are using that method.
@@ordep1132 on what kind of job are you forced to code something while a bunch of people are staring at you, and top of that you have to explain it? The closest thing that comes to mind is a coding instructor. Are you hiring for a coding instructor?
@patrick nah, you haven’t seen popularity. Topcoder was popular. Geeks4geeks couldn’t capitalise on the UI, but still much popular. It just wasn’t the right time for them.
I have to disagree. I can read a good book, but I couldn’t write one. Same thing in your scenario. It needs to be a balance between being able to understand and explain someone else’s code, and writing your own code.
@@pm8rsh233 Sure, it's also important to make sure they can write code too -- but reading someone else's code is a skill that much fewer people seem to have actually learned. And, for the type of software engineering work I do, I spend probably five times as much time digging through open-source to figure out details that aren't in the documentation vs what I spend actually writing code.
@@pm8rsh233I think the person said this because you spend more time debugging, especially code that isn't yours, than really writing code. At least as a junior.
@@hotgoosezion547 And in that case you would need to demonstrate how well you are at problem solving and coding. Being good at reading is only half the requirement.
I did a coding interview that landed me a job. It was not leetcode bs Instead, they gave me some data similar to the one we work with at the company, and asked me to transform it, build an api etc. You know things you actually do do on a daily basis. And they wanted to get a sense of how I go about working on those issues, coming up with the solution, clarify requirements etc
@@ci6516then companies need to stop asking bs questions to developers with decades of experience, because real world problems are what experiences makes us good at.
@@ci6516we do something very similar and it works with juniors too. You have to run tests and fix the existing API. But it’s not testing if you can do it, but how you’ll approach the problem. We had people who would ignore the interviewers. When we give them suggestions they’d say things like “ok” and proceed doing something else. One of the tests is designed so that almost nobody can solve it quickly - it’s an error in an obscure library, and one of the interviewers explains step by step how to fix it but some people repeatedly tell them that they don’t need help or straight out ignore them. We don’t care if you can solve this specific problem yourself but if you can work with people. Just as a side note - we also had somebody who removed failing assertions from the tests and was happy with that solution.
Here in Finland, we have a great way of seeing if a candidate fits with the company and visa versa. It’s called trial period , and usually lasts 3 to 6 months 😂
The moment a company asks me a leetcode style question is the moment I end the interview and feel bad for having wasted my time with a shitty employer. Worked very well so far.
@@fltfathin You can say that about anything though. Leet code problems can be recognized by problem type and the solving algorithms memorized as well -- which is the only way to solve them. Actually figuring out how do the problem isn't usually possible in the time frame they give you, and generally they want an optimized solution, so you *have* to know the solving algorithm already. You aren't going to just "figure out" a solution that some PhD graduate wrote a paper about in 30 minutes.
@@patrickheney9201so true. The brute force and almost optimal (better) solution is easy to crack on your own. But the optimal solution is often a proprietary algorithm which has a name. Only way to know it is to memorize it. U know it the interviewer knows it. But of course u can't say the name directly. So u have to frame it in a way that shows u figured it out on your own. It's mostly a communication test and how u work under pressure
@@patrickheney9201 That's what programming knowledge is all about. Nobody expects everyone to be able to invent novel algorithms on the spot. What leetcode style problems test is that you have knowledge of common solutions to common problems, are able to recognize use cases for those solutions, implement them and explain your thought process.
honestly I've done a ton of interviews, and they all had leet code problems. But I told the interviewee solution does not matter much but let's work on the problem together, so it becomes a back and forth solving problem with a buddy and i just judged if he was cool to work with.
The only good way I have seen leetcode problem being used. My fellow interviewer likes fizz buzz question, but I make it a point that the interviewed candidate should have access to internet for built in methods and syntax checking, and we help out as well.
Imagine the amount of talent that gets forever missed because of the arbitrary leet code puzzle barrier. It literally filters out practically all creative people. What your left with is one track mind people who can get from A to B but only via formulas and they also have giant ego's. Leet code puzzels that have practically no relationship to coding skills will ultimately be the downfall of the companies that use them but as long as they can keep buying out successfull startups, theyll still get some talent coming in the backdoor.
If you’re actually creative and good at what you do, but somehow fail to sit a basic exam, then you’ll find success elsewhere. You’re not immediately shut out of ever coding in your life. It’s not gatekeeping anything.
I imagine that's a very small number. I imagine the venn diagram of people who are very talented but can't complete a simple Leetcode style task has very little overlap. This type of argument just shows that you don't understand what computer science is about. It's not about memorizing "formulas" it's about understanding common solutions to common problems, being able to recognize when a solution is appropriate, implement it and explain your rationale. Most companies don't give the hardest leetcode style problems in interviews, they give fairly simple ones that they expect any competent developer to be able to solve. If you can't solve them then you're probably not a very competent developer.
@@reunionproductions solving leetcode puzzles is just pattern matching memorizations, it has no correlation with real world problem solving. Some of the worst software engineers I know are expert leet code puzzle solvers. They couldn't write a quality peice of software to save their life. Anyone can take time to memorize stuff, it says nothing about their talent as an SE.
@@MyNameIsSalo Leetcode problems are tightly correlated to a few subset of software engineering instead of software engineering as a whole. Being good at this doesn't mean someone is a good software engineer. Just like being good at cooking doesn't mean they are good chefs. It takes a lot more than that.
@reunionproductions You're the misunderstood genius, or is it the person you're interviewing? I've got 38 years in this industry. My career spans the entire history of the internet, starting a year before it was released until now. I've easily got millions of users on the software I wrote, which I'm not going to break down in a RUclips post. See, "genius," what I do is have an intelligent conversation about software. Whether or not they can play "Mary Had a Little Lamb" on LeetCode has no interest to me. I want to see what they're going to build with or for me, not if they can use crayons. LeetCode is how a mid-level, mediocre developer does an interview. I want to talk with the guy personally, make sure he's not an asshole. Lack of skill I can train in a good dude. Being an asshole I can't do anything about, and hiring an asshole can be an expensive mistake.
Hey man sorry to go on with the quote genius that was kind of catty I guess I don't like the words genius because I think most people are just regular people they're not geniuses and I think sometimes we get mistaken for that We're just dudes trying to hold down a job to feed our family and I wish that other developers who are employed would remember that when people are interviewing Hire people you think you can trust and hire people you think you're going to enjoy working with I think its as simple as that
Our union prevents us from getting to technical with the questions we wish to use for applicants. Not because of proprietary stuff, but because they're scared we may stress people out.
As someone with ADHD who is much more articulate once I've had some time to think about things, I'm constantly working on personal projects because I'd rather show an interviewer something I'm building and talk through the problem I'm trying to solve, the scope of the project, the tools I'm using to solve the problem, and how I'm going about it. I stumble over my words constantly in leetcode interviews even though I did a leetcode problem every day for a year. From an ADHD/memory perspective, let me show you something I'm doing that I give a shit about and have been fixating on for 15 months, not something my brain will throw away before the interview is even over.
Bro lol. I feel you. I also obsessed with problems and think about it for long periods at a time. That’s a good trait that you have that separates you from a lot of people. But As someone with ADHD, I get that you are slower cause you get distracted. However I can tell you that in top companies, there are so many people that are QUICK and they also work hard and stay obsessed in the same problem until they find a solution. So unfortunately, you’re competing against people who are both OBSESSED, Hard working, QUICK, and TALENTED. So unfortunately, passion and working hard is not sufficient when you’re competing with more talented people who also have passion and works hard. That’s the reality of the world (At least for top companies).. For smaller unknown companies, people won’t be as talented or at least won’t work as hard.. so it’s possible to shine there
@@kaseywahl It’s related to your comment about how your ADHD makes you suck at leetcode, but that you are passionate about projects. My point is certain companies still care about “leetcode” because it shows how quick you are on your toes, which is important
Don't get me started on HR. I know a lot of people who are good but the HR will reject them based on experience, and the ones I get to interview with 9+ yrs of experience will not know shit once the IDE can't resolve the problem.
@@reunionproductions by solve you mean arrive at an answer? Yes, but I almost always need considerable help. I rarely solve a medium or hard problem without any help
its not sticking because you're not practicing. practicing doesn't mean you move on to the next question after you solve it and stop thinking about it. getting good at leetcode requires the same type of work as getting good at math. you have to do the same problems over and over and over even if you get it right each time. THEN it will stick. And once it sticks, it will be much easier to develop an intuition for similar problems.
We do have a good way of judging whether someone's good at programming within 50 minutes or an hour... It's called giving them an incredibly simple problem to solve that would also require them to ask for requirements and then having them show you the simple solution to it. I was once given a problem that was very simple to solve, but it did require me to ask some questions so I clarify requirements. I, then, use tdd to write a solution. It's not perfect but if you don't know how to ask clarifying questions about the business problem you're trying to solve or your code is just spaghetti when you write it , then you probably aren't good at being a software engineer. At the very least, you're just a junior developer.
Similar reason to why people from top universities who maybe did engineering or a technical field but mot specifically too much programming may be hired because its a filter. “Youre much smarter and a better problem solver than the others, so we (a large company) can afford to polish your skills in a specific technology before you jump into massive responsibilities”
@@cajonesalt0191if this logic had any reason to it this would be true. When people admit the filters they use are crap but don’t come up with better filters you can’t blame the people applying. It’s just a whole lot of corrupt nonsense in every direction no one wants to call out. We hate the bad guy but yet we all work for him and say tough shit if you can’t get in
As coder with 10y experience I think leetcode is a lot easier than describing my past projects and what I did. I just don't remember and I don't prepare..
I personally think that the tests that actually have you build apps using the tooling you need for the job are 100% better and produce more qualified hires. Some of the best devs I've ever met perform poorly in high pressure environments, solving these make believe problems.
I partially agree, but those kinds of tests also have pitfalls: they can be unduly time-consuming for both the candidate and the reviewer, and can demonstrate only superficial proficiencies that don’t apply to difficult problems or different tech stacks (and most tech stacks change eventually). Designing a test which meaningfully reflects the current workload AND demonstrates deeper skills AND doesn’t require hours or days to complete & evaluate can be a tricky needle to thread.
@@nw42 leetcode is the epitome of superficial proficiency, its idea of difficult problems are completely foreign to the difficult problems seen in the real world, and it does nothing to demonstrate knowledge of any tech stack. it doesn't even demonstrate knowledge of a language as all of its problems can be solved with the most basic understanding of a language's standard library.
@@nw42 - Here's some sample data, create an API that does 123. Here are the requirements and desired outcome. Or - Create a mobile app that consumes this public API. Here are the requirements and desired outcome. During the interview: "I took a look at your code can you tell me why you did XYZ instead of ABC? What prompted you to use this library/architecture?" Not that hard.
If someone can solve Leetcode problems one can surely say fhat they are good but just because someone hasn't touched leetcode problem that doesn't mean they are bad. Someone with good rank on Leetcode but no experience with actual real world project will definitely suck, someone who has done real world projects won't struggle much with leetcode problem, but if someone has done both then they are definitely better.
I was doing hiring, and the recruiters would send through candidates for TS devs. "Seniors". I would ask them to do only a couple of things: an array filter and array map. All of them failed. Ridiculous.
@@retagainez I'm pretty sure I'm not wrong about array map and filters. Also pretty sure they could not do it, or at best, got flustered and confused that I would ask such an easy JS question such that they still failed it.
@@ci6516map, filter and even reduce an array is some of the most common thing to do when you want to do data manipulation. It is not something that only reserved for people who study at the university or college, it is taught everywhere so self taught devs are actually more than likely have faced and learned about this.
The problem with most "leet code" problems, is that they rely on the applicant to recognize the type of problem it is and know the solution. It's not about "figuring it out" it's purely about "do you already know what a backpack problem is and the algorithm to solve it, if so, write it down now." Because nobody is going to "figure out" how to do a Knight's Journey in 30 minutes, if they don't already know the algorithm.
Student here and in interviews I only got asked stuff which I explained in pseudocode, to show I get what they are doing. Real code not so much if at all, but its trivial and can be learned pretty fast(syntax etc)
This is like hiring a 3D artist or a texturing artist by asking obsure technical details about the software as opposed to having them walkthrough their portfolio. A good artist who knows their shit will be able to walk you through each design decisions they made, the problems they faced, how they fixed it etc. Likewise someone who knows how to code can easily walk you through their porftfolio, explain why they designed this particular part like this, problems they faced, how they fixed it. So on and so forth. Not to mention 99% of the actual jobs don't have anything like leetcode. Never in my 11 years of programming have I ever had to invert a binary tree of the top of my head. This is such a lazy way to interview candidates. I'd understand using this for a fresh graduate (with limited work experience). But anything else is just pure laziness on the companies part. Also, I have severe adhd, with it comes extremely shor working memory, so I am not good at recalling information off the top of my head always, so whiteboards interviews aren't the best for me. I've also had one interviwer ask me to write up the code to my Final year project and yes he was fucking serious.
That's a great approach, assuming the candidate actually has a portfolio. Leetcode style problems don't have that prerequisite so they're a good "one size fits all" alternative that doesn't expect everyone to have spent all their free time programming for fun. I can understand the frustration some people might feel if they are asked to solve some crazy hard problem and expected to do it in the absolute most optimal way possible. However that is not my experience with these types of interviews. My experience has been fairly simple tasks that I had no problem completing, even in unfamiliar languages. Yet when I do these tasks effortlessly, recruiters get their panties in a bunch like I'm some kind of savant, which in my view just validates the need for this kind of screening. Most of my peers in university sucked at programming. After 3 years they still struggled with things I could do blindfolded in the first year. Personally I support these kinds of tests, because I wouldn't want to hire someone who couldn't complete them and I know for a fact that a degree is absolutely not a guarantee. I also know for a fact that very very few of those peers have any kind of "portfolio" to show. The only reason most of them had a GitHub account was because it was required for certain classes, and the only way they knew how to use Git was by following step by step instructions - and they usually struggled with that as well.
ya, leetcode and other live coding tests really aren’t accommodating whatsoever to people like us who are neurodivergent. Pretty unfair for companies to do this actually. at the very least companies should give you multiple different options to choose from for testing you (live coding, take home etc). whichever you’re most comfortable with.
at my current company, they actually gave us problems to do from home and we had 48 hours to do it. And they did took that long to solve, (this was before gpt btw), they changed things up now
Check the person most recent personal projects and you know how good they are. This and the basic knowledge check about language and framework is more than enough. No new for stupid math or fictional problems. Personally I would never do such crazy tests. I deal with real problems, if a problem isn't real I don't bother with it
People fake or greatly overestimate their contribution to personal/remote projects (this only got worse with the aparition of LLMS from my experience) and frameworks/language methods can be learned by code monkeys. If you want capable problem solvers you will need a little more then that in the screening phase.
@@bogdyeeYou can just ask them a project that build on their own. If they can build the project one their own in several months then this proves they have the motivation, the drive and the knowledge to use the technologies and the tools that they apply into the project. Done, very simple. Even the code monkey will not be able to do something like this, since they don’t have the ability or the urge to research something and applying it into the project in the first place.
@@viethoangtruong54 Its easier to build software on your own. Its much harder to build software in a team. Try contributing to open source and you will know it. There will be days where you are just discussing stuff in github issues or days where your entire PR will be rejected. Personal projects are not a good indicator of skill unless you have built something substantial thats used by many(think automatic1111 etc) or contributed to a big project like kafka, linux, etc. No one cares about your todo app or some fake social network site with no users.
I've noped out of some nasty online tests along the lines of "find the size of all squares in a matrix representing a canvas, without googling anything". Not that that isn't a useful thing to figure out but that's a big one to just solve off the top of your head, though I guess it depends on the domain and the level youre going to work in.
I think thats the answer right there - hiring is hard, and 30, 60, or 120 minutes is not enogh to assess how good someone is. It’s like judging someone you will marry after the 1st date. The shit comes out in a couple of months.
Yeah, but it really sucks to hire someone only to have to get rid of them two months later... same as it sucks to get hired, quit your current job, then lose your new job two months later. So anything that can give a good signal that someone will probably still look good in two months, but in a much shorter time, is useful.
@@BittermanAndy of course, both sides lose time and money, nobody wants this to happen. Unfortunately that's just what it is - it's first impressions, which can be very deceiving (make bad ppl look good, and good ppl bad).
I think at least in Europe, being good at LeetCode nearly always means you’re at least decent at coding and will be a a good addition to most development teams.
Honestly when I interview people most of the time the questions are just there to provide some kind of structure. My goal is to talk shop with the candidate and get a sense of how much they know, how they approach a problem, if there's any red flags in how they behave, etc. I actually like to give candidates some questions that are frankly BS which I don't expect most people will get right, for two reasons: 1. if they _do_ get it right, I'm impressed and 2. it shows how they handle difficult situations under pressure, which is a more important skill than usual at my company since so much of the work we do is research-oriented. When we start a project, we usually don't know how to solve it, and technically we don't even know if we _can_ solve it. If I ask the candidate some difficult question and they complain, whine, act like they're above answering the question, I already know they won't be a good fit.
I just recently got into the whole leetcode thing. I still don’t have a good grasp of most of the fundamentals of arrays and tree nodes and stuff but I’ve done a lot of projects that have given me a solid understanding of core web development principles and how to approach complex problems. Sadly tho, seems like if I want to get a job as a software engineer, I need to learn how to leetcode properly.
@@arifrahim7247 To be clear, we don't do anything to do with webdev, and it's not really an industry I understand. My interview sensibilities are shaped by the R&D sort of work that we do.
I might be an exception, but I did a lot of trees and graphs type work in my first year at big tech. You can’t really learn on the job or google search when you need to problem solve much harder problems than any hard level type leetcode problem. You also need to work collaboratively with others who know their fundamentals and some who are inventing new DS&A. It was amazingly fun, but hard so I’m glad I went through leetcode.
Arrays, maps, sets, and trees are all versions of arrays. I think it's the algorithms leetcode teaches you. Which you may never use again. They're all for a specific use case. That you can look up for a specific problem.
I like leetcoding, it has definitely improved my coding skills. Even if you have been writing code for many years, sometimes writing boilerplate code will not help you grow ability to write code creatively. I think doing leetcode type questions helps with that.
Do you consider LC the best platform to focus your energy on, or do you also use others like HackerRank, Exercism, etc.? I've never used any of them TBH, but I think it'd help me do better in technical interviews, even after 12 years in industry.
That does make sense, however in this current job market they are expecting you complete the problem 100% Also now these companies use third party companies to do these technical interviews and ya if you don't complete 100% in the allocated time, you don't go through to the actual engineers with "the secret handshake". In my experience of only a month of job hunting, so far, and studying these leet code stuff, I am just really bad at these word problems but thankfully I know how to build a real project to keep my motivation going. Thank you for this bit of knowledge though, I am going to use it going forward for these interviews, cause sadly, they are all needing this 100% complete handshake.
It’s kinda how everyone I’ve met that uses Vim is a great programmer. Not because Vim makes them great, but because the kind of person that takes that much time to get good at it will probably put the same effort into everything else. Someone who’s good at coding challenges is likely going to put it work in every other skill
How about VS Code Vim like me. I love V motion but wont setup for my project. Altough i felt like superior when editing config through ssh when looking down nano users
Because a portfolio of actual development job experience or contributions to a code base clearly doesn’t matter in *checks notes* a Software Engineering position. Instead, let’s have you solve problems that you’ll never even tangentially interact with and that have been optimized for years?
I've been hiring software engineers for the past 10 years. I have a better method. Give them some dirty code and ask them to clean it up and make it more object oriented. This reflects 90% of the work they'll be doing day to day. If they can effectively do this they're hired.
I hear this a lot that companies do this to see if you're good enough to do the job but being good at interviewing doesn't show you are good at the day to day work, or working in a team, or dealing with tight deadlines so isn't it still a risk either way? I do/did the handshake anyways because it is what it is at this point and I was hungry to change my situation but saying this handshake is necessary to know I have a good pick isn't really true right? I wonder if a probationary period, similar to an internship, is a better method. Idk a better way either, but I don't think the handshak is the best way to weed out the best candidates. I also wonder what other methods have been tried in the past🤔... I know some companies are trying different options now, which is really good, and I hope it can lead to positive findings.
Imagine showing up to the hospital for an interview, and they say, " in order for me to know if you are really good doctor, please operate on this patient, in the next 20 minutes"😂😂😂😂
Yes the secret handshake is stupid. Is like nepotism. I refuse to do leet code problems because they dont represent reality. I can waste 1 month learning all the answers and then after i get the job i am exhausted from all that learning. A lazy interviewer will use leetcode problems. A good one will find a few questions personalized for each candidate that shows how good they are. When i need a team member that just have to do repetitive tidious work i might use leetcode. When i need problem solvers i run from that shit. And if you have tidious repetitive work to do you might just as well get someone with problem solving skills to automate that process.
I see two reasons: - A short interview requires some kind of pass/fail task, even if it isn't particularly well correlated with the job. - Nerds can't help but gameify everything.
Im a couple of years into my career and work w the NOAA as of recently. I've never used leet code more than a few times. I consider myself very capable and able to tackle any problems thrown at me. However, I I hesitate to apply to a FAANG company because I feel like I would bomb the leet code portion. I use so many tech stacks and frameworks and languages, It's hard to keep track of things like syntax and such, and it worries me. On the other hand, because of that, I'm not afraid to learn new languages and tackle challenging problems because I can recognize the similarities in all of these languages and I feel like it makes me a really strong candidate.
90% of the time, you are going to be hiring someone who hasn't done what you want them to do. So instead, companies need to figure out if the person can do what they want them to. So adaptability and ability to start to approach and solve problems is the thing they'll try to test. And that is not easy.
Since the questions can sometimes be randomly selected would it ever be okay to ask for another problem if they give you a problem statement and you don’t know that one? Just say “here’s my initial thought but getting into into this I am not quite sure about going down this path. Do you have any other questions prepared?”
you can take this knowledge and shove it. i have 20+ years in development and all i want to know during the interview, is whether i can reason with you. i would never hire this guy, cause he loves the sound of his voice too much.
Sorting algorithms is simply just working with numbers in a set that fit into an ordering different sorting algorithms might have different orderings. Memorizing the syntax of how to code them is different from understanding what merge sort has to do with total ordering and how this can be abstracted in the first place. LeetCode doesn't teach this.
My company gave me an assignment. I had a week to build a PWA shopping app in accordance with a predefined design and integrate it with an existing API. So while there is no good way of gauging how good someone is at coding in 50 minutes, who says the process needs to be confined to one interview? No offence, but the only companies who've ever given me leet code interviews are shitty companies I've walked out of.
I hate take home assignments like yours. After spending an entire week they will simply reject without any feedback nor will give any monetary compensation for the time I put in. Its much better to do a "standardized" leetcode interview in your language of choice. Like for example if you have never worked with java techstack, you won't even be considered for takehome styled interview, however with leetcode you can choose any language and give the interview regardless of if the company uses java or any other language/tech stack
I prefer problems that give the candidate multiple sets of data that has some relationship and then ask them to solve various problems by merging, joining and filtering.
The problem with leetcode questions is that they appear to be good if you're a bad interviewer, but are not actually good interview questions. So they have become a common trap bad interviewers/tech companies fall into, like something a junior interviewer would do while a senior would know better.
I've NEVER done Leetcode for an interview and never will. LC is like asking a mechanic about the mathematical equations for the physics for internal combustion. EDIT: you know what I meant. Leetcode is NOT what the job entails. It is pretty much pointless. When was the last time you inverted a binary tree at work?
100% agree. Are we going to be only working on random questions everyday? such as: "Write an algorithm that figures out how many bananas are being eaten by 10 year olds who are watching TikToks." ... didn't think so.
LC just doesn’t yield practical experience. At least the 5 year react dev that doesn’t know array functions does know how to use 3rd party libraries, and that’s a better skill most of the time at least unless you’re making the 3rd party library to begin with, i suppose.
I've been coding for 20+ years. I've coded, led teams, architected solutions, hired teams, all of my projects have been successful... and I've never even heard of Leetcode.
I've seen professional students in a number of different fields, and they're the worst to work with. They think they know stuff, but can't actually do the job that's asked of them when they need to solve a real problem, and they lack the analytics skills to evaluate why one solution might be better than another. I feel like LC caters to people who study for rote memorization without the thinking skills needed for the real world. I want someone who understands the problem, someone who can provide a solution someone else might not have thought of... Not someone who gauged all their answers by what was in the back of someone else's book.
But if you can't build Spotify type app BUT you can do a million Leetcode problems. Why are you being hired? I find knowing how to put all this stuff together into something working is a skill on it's own.
They want to hear me talk? Over the online-with-no-intervewer-leet-code-challenge technical interview? Damn, should have known that I had to record my voice and send them an email with the file afterward...
ok this actually changed my mind about these questions. I always thought of them as just iq tests, essentially, except ones that are easy to cheat on and not indicative of much wrt how you'll be as an employee (except insofar as intelligence helps with everything). but testing for hard work and preparation kinda makes it seem more reasonable, especially as a way to counterbalance intelligence - you have to be either very smart, very hardworking, or somewhere along the function between the two that allows you to compensate with one for lack of the other.
I think leetcode is a great complementary skill to any programmers. it strengthen problem-solving skills and proficiency in DSA concepts like array, sets, maps. I think those with good DSA skills can learn new tools pretty quickly. I learned a bit of React after 2 days and could already build something look somewhat good (but the code and folder structure behind is disgusting lol, lots of DSA-ing gave me the mind power to not be afraid of potentially typing hundreds lines of code)
why not demonstrate your ability to learn new tools quickly by having multiple projects that use different tech stacks that you can talk about in-depth?
I never ask leetcode questions. I asked scenario based questions based on what they have written on their resume. My favourite one is to write a class with properties that can store data from an xml file. Simple question that test both knowledge and problem solving.
Leetcode also scales from the company's POV. If you have 50 people interviewing 200 engineers, for 10 positions, striving for a consistent grading process for all 200 interviews is really important. Way more consistent than "tell me about what you've built" at least.
“I’ve built nothing but LC problems, filling in those functions are my apps, I’ve never used an external library in my life, I always recreate the wheel”, yeah… that’s real trust inspiring…
The only consistent thing you get out of doing that is you can happily throw away a lot of those applications. That's what Leetcode is REALLY doing, giving people a convenient way to narrow shit down.
I think if you are able to gring 300 lc problems. You can learn anything about the job. Because grinding that is extemly boring so at a real job it wont be that bad
Our local big techs say "you have to be able to code without bugs so that you spend less time debugging when you write our production code". I call BS.
I've chosen not to hire people who have all the answers without thinking about it before, drag that handshake on to make it seem like you're solving it in real time vs reiterating the answer
If you can afford it then take home interview projects are the best way to guage their knowledge. With take home projects you actually go beyond just solving random abstract questions which are usually found on leet code and put the interviewee at the test for the things they're actually going to work for years if they get selected. Some important things which are just hard to guage from a pure LC interview - Code quality, version control, testing, practical problem solving.
I don’t think they need to be take home challenges. You can present a candidate with a simpler issue that can be solved, or at least have them explain how they would go about solving it. The challenge can be representative of what you do in the job, you get a notion of how the candidate thinks and communicates, that’s a lot more valuable than memorizing some algorithms for LC. Plus it respects their time I wouldn’t expect a candidate to dedicate multiple hours of their own time to solve a problem when chances are they’re interviewing for other jobs as well
I disagree with this one. Much better to show them some of your system's code, and have them describe what it might be doing and why it might be doing it that way. Forget the leetCode test garbage.
@redfordkobayashi6936 Even leetcode easy requires things that many senior devs haven't ran into. If you ask someone to do an Inorder Traversal of a binary tree, the chances that they ran into a single problem that required them to write a binary tree in their day to day life programming professionally is pretty low. And for good reason. A binary tree is probably not the best solution for a lot of their day to day problems. So they're going to be stuck on what "Inorder traversal" even means for a bit. Of course, this all depends on the job you're applying for. If you're applying for a job at a company in which such data structures are common place then yeah you should be expected to understand them. But even then, just showing your actual code to the person and having them describe what it's doing? Accomplishes both. If such data structures *aren't* needed. Then your sample won't have them. If such data structures *are* needed. Then your sample will probably have them. Is this possibly more difficult? Yes. But it's solving the core problem of the question being relevant to the actual job.
The problem is a very high number of poor candidates who cannot even solve a simple problem. Use LC or similar for screening and then pair programming for the actual interview.
The reality is that hard work is what makes most people good at a job. If you've grinded 300 leetcode problems, you have, at the bare minimum, established that you are have some work ethic to you. People don't like it, but showing up and giving consistent effort over time puts you above a significant % of other candidates.
If you can spend several months alone to build a project from the ground up then for me that is much more impressive that any grind you can do on leetcode.
maps are just arrays that dont confine their indexes to integers only, you can use strings and such to store variables, think [ 'price' => 123, 'description' => 'zxcxc', 'sometingElse' => [ other map here ] ] instead of [123, 'zczcx', [ other array here ] ]
I usually ghost companies when they try and give me tests or assignments. Maybe if I was unemployed I'd have time for that but I'd rather go outside after work and do non work related shit.
The thing I don't get about this, is that these companies are highering programmers who will work for them for months or maybe even years. Doesn't that mean it would be worth whatever time it needs to interview a person in a more representative setting? This could mean asking them about how they handled past projects and maybe give them a more complex problem with the required time to solve it like one would on the actual job. The amount of time you waste with programmers that might have been better on the leet code tests than the actual job or even worse the talented people you loose to other companies should be a major factor when coming up with an interview process.
And then there is that person who ONLY studied leetcode problems, but can't solve any real problems.
Unfortunately true.
Which is most juniors
They eventually learn,
At least they built analytical thinking and problem solving skills by solving those problems
because they are junior
It's just a baseline, to filter out applicants who can't do the "secret handshake". Ppl will still need to be interviewed and can be trained afterwards
Hiring an architect based on their still at jenga basically.
😂😂😂😂 that makes it hilarious and depressing and now I don't know if to trust the industry 😂😂😂😂
It's true
"So how'd you make the building even taller? Thought you had limited resources"
"Oh we just took the stuff from below and put it at top"
"You did what-"
@@johnnycripplestar5167 There's a building like that in some rich people's gated park (can't remember what YT vid I saw it in)
if you aint good at jenga I aint hiring you to build a building.
I ask them to talk about something they built in the past. I get into the details, challenges, and choices. It becomes obvious pretty quickly if they understand what they did, rather than just put some keyword on a resume.
I do the same. This “leetcode” nonsense has to stop. I’d rather see the person’s eyes light up talking about their project and how engaged they are when we “what if” the use case or their assumptions.
Companies give 6 round leetcode interviews for their mundane “make a corporate dashboard” level jobs. Absolutely counterproductive.
Same, that's exactly how I hire all my devs. I also volunteer at 2 code camps doing mock interviews, and it is super obvious who the best coders are using that method.
I don’t agree. Doing this leet code style live programming is a great way to evaluate how someone communicates and solves problems.
@@ordep1132disagree, it requires memorization and that’s all.
@@ordep1132 on what kind of job are you forced to code something while a bunch of people are staring at you, and top of that you have to explain it? The closest thing that comes to mind is a coding instructor. Are you hiring for a coding instructor?
It's crazy how algorithms got a brand name, leetcode
It's a popular website for coding challenges. It's not just a random "brand name" for "algorithms."
@@patrickheney9201 sarcasm bro
Imagine having to work with someone primarily over Slack who can't recognize humor
@patrick nah, you haven’t seen popularity. Topcoder was popular. Geeks4geeks couldn’t capitalise on the UI, but still much popular. It just wasn’t the right time for them.
@vinayaktripathi8307 that's how you know he is actually good just not at parties
I've found, the real good interview isn't "write a code that solves X", it's "here's some code, explain what it does and why"
I have to disagree. I can read a good book, but I couldn’t write one. Same thing in your scenario.
It needs to be a balance between being able to understand and explain someone else’s code, and writing your own code.
@@pm8rsh233 Sure, it's also important to make sure they can write code too -- but reading someone else's code is a skill that much fewer people seem to have actually learned. And, for the type of software engineering work I do, I spend probably five times as much time digging through open-source to figure out details that aren't in the documentation vs what I spend actually writing code.
@@pm8rsh233I think the person said this because you spend more time debugging, especially code that isn't yours, than really writing code. At least as a junior.
@@hotgoosezion547 And in that case you would need to demonstrate how well you are at problem solving and coding. Being good at reading is only half the requirement.
If you're interviewing someone to fill a QA role then this would be an excellent test
I did a coding interview that landed me a job. It was not leetcode bs Instead, they gave me some data similar to the one we work with at the company, and asked me to transform it, build an api etc. You know things you actually do do on a daily basis. And they wanted to get a sense of how I go about working on those issues, coming up with the solution, clarify requirements etc
That’s senior stuff not junior
@@ci6516then companies need to stop asking bs questions to developers with decades of experience, because real world problems are what experiences makes us good at.
@@ci6516we do something very similar and it works with juniors too. You have to run tests and fix the existing API.
But it’s not testing if you can do it, but how you’ll approach the problem. We had people who would ignore the interviewers. When we give them suggestions they’d say things like “ok” and proceed doing something else.
One of the tests is designed so that almost nobody can solve it quickly - it’s an error in an obscure library, and one of the interviewers explains step by step how to fix it but some people repeatedly tell them that they don’t need help or straight out ignore them.
We don’t care if you can solve this specific problem yourself but if you can work with people.
Just as a side note - we also had somebody who removed failing assertions from the tests and was happy with that solution.
@@ci6516this is junior stuff at big tech companies
You can't do that when interviewing someone fresh out of college
Here in Finland, we have a great way of seeing if a candidate fits with the company and visa versa. It’s called trial period , and usually lasts 3 to 6 months 😂
This'd be the best but how can you apply trial period on 100 candidates?
The moment a company asks me a leetcode style question is the moment I end the interview and feel bad for having wasted my time with a shitty employer. Worked very well so far.
Its almost as if portfolio projects dont matter anymore 😂
Sadly portofolio can be faked or copied over so nyeeh
@@fltfathinand leetcode solutions can be Googled
@@fltfathin You can say that about anything though. Leet code problems can be recognized by problem type and the solving algorithms memorized as well -- which is the only way to solve them. Actually figuring out how do the problem isn't usually possible in the time frame they give you, and generally they want an optimized solution, so you *have* to know the solving algorithm already. You aren't going to just "figure out" a solution that some PhD graduate wrote a paper about in 30 minutes.
@@patrickheney9201so true. The brute force and almost optimal (better) solution is easy to crack on your own. But the optimal solution is often a proprietary algorithm which has a name. Only way to know it is to memorize it. U know it the interviewer knows it. But of course u can't say the name directly. So u have to frame it in a way that shows u figured it out on your own. It's mostly a communication test and how u work under pressure
@@patrickheney9201 That's what programming knowledge is all about. Nobody expects everyone to be able to invent novel algorithms on the spot. What leetcode style problems test is that you have knowledge of common solutions to common problems, are able to recognize use cases for those solutions, implement them and explain your thought process.
"...map, for, filter..."
[reduce has entered the chat. Feeling hurt.]
He left me scratching my head there.
Reduce is for hipsters. Just accumulate over a forEach like the rest of us
defining map in terms of reduce: 🗿
[flatMap, every, forEach, some, reduceRight] : ....
@@icarus212121Yeah, do that in any fp language...
honestly I've done a ton of interviews, and they all had leet code problems. But I told the interviewee solution does not matter much but let's work on the problem together, so it becomes a back and forth solving problem with a buddy and i just judged if he was cool to work with.
The only good way I have seen leetcode problem being used.
My fellow interviewer likes fizz buzz question, but I make it a point that the interviewed candidate should have access to internet for built in methods and syntax checking, and we help out as well.
Imagine the amount of talent that gets forever missed because of the arbitrary leet code puzzle barrier. It literally filters out practically all creative people. What your left with is one track mind people who can get from A to B but only via formulas and they also have giant ego's. Leet code puzzels that have practically no relationship to coding skills will ultimately be the downfall of the companies that use them but as long as they can keep buying out successfull startups, theyll still get some talent coming in the backdoor.
If you’re actually creative and good at what you do, but somehow fail to sit a basic exam, then you’ll find success elsewhere. You’re not immediately shut out of ever coding in your life. It’s not gatekeeping anything.
I imagine that's a very small number. I imagine the venn diagram of people who are very talented but can't complete a simple Leetcode style task has very little overlap.
This type of argument just shows that you don't understand what computer science is about. It's not about memorizing "formulas" it's about understanding common solutions to common problems, being able to recognize when a solution is appropriate, implement it and explain your rationale.
Most companies don't give the hardest leetcode style problems in interviews, they give fairly simple ones that they expect any competent developer to be able to solve. If you can't solve them then you're probably not a very competent developer.
@@reunionproductions solving leetcode puzzles is just pattern matching memorizations, it has no correlation with real world problem solving. Some of the worst software engineers I know are expert leet code puzzle solvers. They couldn't write a quality peice of software to save their life. Anyone can take time to memorize stuff, it says nothing about their talent as an SE.
@xpload32004 you speak from experience, where do you get that they couldn't write any piece of software?
@@MyNameIsSalo Leetcode problems are tightly correlated to a few subset of software engineering instead of software engineering as a whole. Being good at this doesn't mean someone is a good software engineer. Just like being good at cooking doesn't mean they are good chefs. It takes a lot more than that.
Here's a reason. They use leetcode questions because 90% of the interviewers can't think of an actual question themselves.
@reunionproductions You're the misunderstood genius, or is it the person you're interviewing?
I've got 38 years in this industry. My career spans the entire history of the internet, starting a year before it was released until now.
I've easily got millions of users on the software I wrote, which I'm not going to break down in a RUclips post.
See, "genius," what I do is have an intelligent conversation about software. Whether or not they can play "Mary Had a Little Lamb" on LeetCode has no interest to me. I want to see what they're going to build with or for me, not if they can use crayons.
LeetCode is how a mid-level, mediocre developer does an interview.
I want to talk with the guy personally, make sure he's not an asshole. Lack of skill I can train in a good dude. Being an asshole I can't do anything about, and hiring an asshole can be an expensive mistake.
Hey man sorry to go on with the quote genius that was kind of catty I guess I don't like the words genius because I think most people are just regular people they're not geniuses and I think sometimes we get mistaken for that
We're just dudes trying to hold down a job to feed our family and I wish that other developers who are employed would remember that when people are interviewing
Hire people you think you can trust and hire people you think you're going to enjoy working with I think its as simple as that
Our union prevents us from getting to technical with the questions we wish to use for applicants. Not because of proprietary stuff, but because they're scared we may stress people out.
As someone with ADHD who is much more articulate once I've had some time to think about things, I'm constantly working on personal projects because I'd rather show an interviewer something I'm building and talk through the problem I'm trying to solve, the scope of the project, the tools I'm using to solve the problem, and how I'm going about it.
I stumble over my words constantly in leetcode interviews even though I did a leetcode problem every day for a year.
From an ADHD/memory perspective, let me show you something I'm doing that I give a shit about and have been fixating on for 15 months, not something my brain will throw away before the interview is even over.
Lol, I probably have ADHD and I prefer doing leetcode than big projects. You have to be disciplined when you build something big.
Bro lol. I feel you. I also obsessed with problems and think about it for long periods at a time. That’s a good trait that you have that separates you from a lot of people. But
As someone with ADHD, I get that you are slower cause you get distracted. However I can tell you that in top companies, there are so many people that are QUICK and they also work hard and stay obsessed in the same problem until they find a solution.
So unfortunately, you’re competing against people who are both OBSESSED, Hard working, QUICK, and TALENTED.
So unfortunately, passion and working hard is not sufficient when you’re competing with more talented people who also have passion and works hard. That’s the reality of the world
(At least for top companies)..
For smaller unknown companies, people won’t be as talented or at least won’t work as hard.. so it’s possible to shine there
@@aznstride4325 I don't know what you're responding to, but it isn't my comment.
@@kaseywahl It’s related to your comment about how your ADHD makes you suck at leetcode, but that you are passionate about projects.
My point is certain companies still care about “leetcode” because it shows how quick you are on your toes, which is important
I have ADHD too (diagnosed October 2023) and I've legit just been brute forcing front end into my brain since December learning by building projects
So happy to see the comments being full of support for properly structured processes and not this stupid letcode like challenges....🎉
Unfortunately, HR does not work this way.
😂😂😂 stupid leetcode questions.....
Don't get me started on HR. I know a lot of people who are good but the HR will reject them based on experience, and the ones I get to interview with 9+ yrs of experience will not know shit once the IDE can't resolve the problem.
Bro Leetcode is the only way of generalization , everything else can be taught , or can be learnt easily .
I keep practicing Leetcode but never feel like i learn anything. Like it doesn't stick
That's because a lot of it is memorization of boilerplate code
@@reunionproductions by solve you mean arrive at an answer? Yes, but I almost always need considerable help. I rarely solve a medium or hard problem without any help
its not sticking because you're not practicing. practicing doesn't mean you move on to the next question after you solve it and stop thinking about it. getting good at leetcode requires the same type of work as getting good at math. you have to do the same problems over and over and over even if you get it right each time. THEN it will stick. And once it sticks, it will be much easier to develop an intuition for similar problems.
@greekthejimmy4107 so do you recommend essentially just grinding a list like Blind 75 until I've memorized the problems and solutions?
exactly how i feel hahaha
We do have a good way of judging whether someone's good at programming within 50 minutes or an hour... It's called giving them an incredibly simple problem to solve that would also require them to ask for requirements and then having them show you the simple solution to it.
I was once given a problem that was very simple to solve, but it did require me to ask some questions so I clarify requirements. I, then, use tdd to write a solution. It's not perfect but if you don't know how to ask clarifying questions about the business problem you're trying to solve or your code is just spaghetti when you write it , then you probably aren't good at being a software engineer. At the very least, you're just a junior developer.
Similar reason to why people from top universities who maybe did engineering or a technical field but mot specifically too much programming may be hired because its a filter. “Youre much smarter and a better problem solver than the others, so we (a large company) can afford to polish your skills in a specific technology before you jump into massive responsibilities”
worst thing imo. thats how you supress talents...
@@duckydude20 If they were that talented, they'd get hired.
@@cajonesalt0191if this logic had any reason to it this would be true. When people admit the filters they use are crap but don’t come up with better filters you can’t blame the people applying. It’s just a whole lot of corrupt nonsense in every direction no one wants to call out. We hate the bad guy but yet we all work for him and say tough shit if you can’t get in
I don't understand anything this guy is talking about, yet I listened to the entire thing.
As coder with 10y experience I think leetcode is a lot easier than describing my past projects and what I did. I just don't remember and I don't prepare..
I personally think that the tests that actually have you build apps using the tooling you need for the job are 100% better and produce more qualified hires.
Some of the best devs I've ever met perform poorly in high pressure environments, solving these make believe problems.
I partially agree, but those kinds of tests also have pitfalls: they can be unduly time-consuming for both the candidate and the reviewer, and can demonstrate only superficial proficiencies that don’t apply to difficult problems or different tech stacks (and most tech stacks change eventually). Designing a test which meaningfully reflects the current workload AND demonstrates deeper skills AND doesn’t require hours or days to complete & evaluate can be a tricky needle to thread.
@@nw42 leetcode is the epitome of superficial proficiency, its idea of difficult problems are completely foreign to the difficult problems seen in the real world, and it does nothing to demonstrate knowledge of any tech stack. it doesn't even demonstrate knowledge of a language as all of its problems can be solved with the most basic understanding of a language's standard library.
@@nw42
- Here's some sample data, create an API that does 123. Here are the requirements and desired outcome.
Or
- Create a mobile app that consumes this public API. Here are the requirements and desired outcome.
During the interview:
"I took a look at your code can you tell me why you did XYZ instead of ABC? What prompted you to use this library/architecture?"
Not that hard.
If someone can solve Leetcode problems one can surely say fhat they are good but just because someone hasn't touched leetcode problem that doesn't mean they are bad. Someone with good rank on Leetcode but no experience with actual real world project will definitely suck, someone who has done real world projects won't struggle much with leetcode problem, but if someone has done both then they are definitely better.
It’s why I went to Uni. I did it, so you know I can slog through stuff even if I don’t love it.
Lol, waste of time and money for things that pretty much basic
That’s a pretty expensive lesson lol
@@jagi7976Not necessarily, not every country will set you back a million to attend uni
@@valizeth4073 in the US and Canada, a lot of Asian and most third world countries it does. I guess not if you’re in Europe
@@jagi7976 correct, europe doesnt debt trap students.
Wow, bro unintentionally summarized the entire education system
😂
I was doing hiring, and the recruiters would send through candidates for TS devs. "Seniors". I would ask them to do only a couple of things: an array filter and array map. All of them failed. Ridiculous.
Self taught devs
What's more likely? You being wrong or everyone else being wrong?
@@retagainez I'm pretty sure I'm not wrong about array map and filters. Also pretty sure they could not do it, or at best, got flustered and confused that I would ask such an easy JS question such that they still failed it.
@@ci6516Self taught often means they have a passion or doing it
@@ci6516map, filter and even reduce an array is some of the most common thing to do when you want to do data manipulation. It is not something that only reserved for people who study at the university or college, it is taught everywhere so self taught devs are actually more than likely have faced and learned about this.
The problem with most "leet code" problems, is that they rely on the applicant to recognize the type of problem it is and know the solution. It's not about "figuring it out" it's purely about "do you already know what a backpack problem is and the algorithm to solve it, if so, write it down now." Because nobody is going to "figure out" how to do a Knight's Journey in 30 minutes, if they don't already know the algorithm.
Student here and in interviews I only got asked stuff which I explained in pseudocode, to show I get what they are doing.
Real code not so much if at all, but its trivial and can be learned pretty fast(syntax etc)
This is like hiring a 3D artist or a texturing artist by asking obsure technical details about the software as opposed to having them walkthrough their portfolio. A good artist who knows their shit will be able to walk you through each design decisions they made, the problems they faced, how they fixed it etc. Likewise someone who knows how to code can easily walk you through their porftfolio, explain why they designed this particular part like this, problems they faced, how they fixed it. So on and so forth. Not to mention 99% of the actual jobs don't have anything like leetcode. Never in my 11 years of programming have I ever had to invert a binary tree of the top of my head. This is such a lazy way to interview candidates. I'd understand using this for a fresh graduate (with limited work experience). But anything else is just pure laziness on the companies part.
Also, I have severe adhd, with it comes extremely shor working memory, so I am not good at recalling information off the top of my head always, so whiteboards interviews aren't the best for me. I've also had one interviwer ask me to write up the code to my Final year project and yes he was fucking serious.
That's a great approach, assuming the candidate actually has a portfolio. Leetcode style problems don't have that prerequisite so they're a good "one size fits all" alternative that doesn't expect everyone to have spent all their free time programming for fun.
I can understand the frustration some people might feel if they are asked to solve some crazy hard problem and expected to do it in the absolute most optimal way possible. However that is not my experience with these types of interviews. My experience has been fairly simple tasks that I had no problem completing, even in unfamiliar languages. Yet when I do these tasks effortlessly, recruiters get their panties in a bunch like I'm some kind of savant, which in my view just validates the need for this kind of screening.
Most of my peers in university sucked at programming. After 3 years they still struggled with things I could do blindfolded in the first year. Personally I support these kinds of tests, because I wouldn't want to hire someone who couldn't complete them and I know for a fact that a degree is absolutely not a guarantee. I also know for a fact that very very few of those peers have any kind of "portfolio" to show. The only reason most of them had a GitHub account was because it was required for certain classes, and the only way they knew how to use Git was by following step by step instructions - and they usually struggled with that as well.
ya, leetcode and other live coding tests really aren’t accommodating whatsoever to people like us who are neurodivergent. Pretty unfair for companies to do this actually. at the very least companies should give you multiple different options to choose from for testing you (live coding, take home etc). whichever you’re most comfortable with.
at my current company, they actually gave us problems to do from home and we had 48 hours to do it. And they did took that long to solve, (this was before gpt btw), they changed things up now
I've done only around 20 leetcode problems because they are too easy, but I am an international master in codeforces, does that work? :p
Works well with tech giants with thousands of applications but when every other company starts doing it this becomes ridiculous.
Check the person most recent personal projects and you know how good they are. This and the basic knowledge check about language and framework is more than enough. No new for stupid math or fictional problems. Personally I would never do such crazy tests. I deal with real problems, if a problem isn't real I don't bother with it
People fake or greatly overestimate their contribution to personal/remote projects (this only got worse with the aparition of LLMS from my experience) and frameworks/language methods can be learned by code monkeys. If you want capable problem solvers you will need a little more then that in the screening phase.
@@bogdyeeYou can just ask them a project that build on their own. If they can build the project one their own in several months then this proves they have the motivation, the drive and the knowledge to use the technologies and the tools that they apply into the project. Done, very simple. Even the code monkey will not be able to do something like this, since they don’t have the ability or the urge to research something and applying it into the project in the first place.
@@viethoangtruong54 Its easier to build software on your own. Its much harder to build software in a team. Try contributing to open source and you will know it. There will be days where you are just discussing stuff in github issues or days where your entire PR will be rejected. Personal projects are not a good indicator of skill unless you have built something substantial thats used by many(think automatic1111 etc) or contributed to a big project like kafka, linux, etc. No one cares about your todo app or some fake social network site with no users.
I've noped out of some nasty online tests along the lines of "find the size of all squares in a matrix representing a canvas, without googling anything". Not that that isn't a useful thing to figure out but that's a big one to just solve off the top of your head, though I guess it depends on the domain and the level youre going to work in.
I think thats the answer right there - hiring is hard, and 30, 60, or 120 minutes is not enogh to assess how good someone is. It’s like judging someone you will marry after the 1st date. The shit comes out in a couple of months.
Yeah, but it really sucks to hire someone only to have to get rid of them two months later... same as it sucks to get hired, quit your current job, then lose your new job two months later.
So anything that can give a good signal that someone will probably still look good in two months, but in a much shorter time, is useful.
@@BittermanAndy of course, both sides lose time and money, nobody wants this to happen. Unfortunately that's just what it is - it's first impressions, which can be very deceiving (make bad ppl look good, and good ppl bad).
I think at least in Europe, being good at LeetCode nearly always means you’re at least decent at coding and will be a a good addition to most development teams.
Honestly when I interview people most of the time the questions are just there to provide some kind of structure. My goal is to talk shop with the candidate and get a sense of how much they know, how they approach a problem, if there's any red flags in how they behave, etc. I actually like to give candidates some questions that are frankly BS which I don't expect most people will get right, for two reasons: 1. if they _do_ get it right, I'm impressed and 2. it shows how they handle difficult situations under pressure, which is a more important skill than usual at my company since so much of the work we do is research-oriented. When we start a project, we usually don't know how to solve it, and technically we don't even know if we _can_ solve it. If I ask the candidate some difficult question and they complain, whine, act like they're above answering the question, I already know they won't be a good fit.
I just recently got into the whole leetcode thing. I still don’t have a good grasp of most of the fundamentals of arrays and tree nodes and stuff but I’ve done a lot of projects that have given me a solid understanding of core web development principles and how to approach complex problems. Sadly tho, seems like if I want to get a job as a software engineer, I need to learn how to leetcode properly.
@@arifrahim7247 To be clear, we don't do anything to do with webdev, and it's not really an industry I understand. My interview sensibilities are shaped by the R&D sort of work that we do.
I might be an exception, but I did a lot of trees and graphs type work in my first year at big tech. You can’t really learn on the job or google search when you need to problem solve much harder problems than any hard level type leetcode problem. You also need to work collaboratively with others who know their fundamentals and some who are inventing new DS&A. It was amazingly fun, but hard so I’m glad I went through leetcode.
Such a good way to talk about this. So acurate
A degree used to be that secret handshake.
I refuse to interview at a company that has this style of interviewing.
Will prime start with the Thor strategy? That would be great btw
I have been in discussions with my editor on doing shorts now for about 2 months. Thor gave me the kick in the butt to do it
@@ThePrimeTimeagen amazing, thor is going to the moon already
This is the exact same secret handshake that some companies are doing for certifications.
Arrays, maps, sets, and trees are all versions of arrays. I think it's the algorithms leetcode teaches you. Which you may never use again. They're all for a specific use case. That you can look up for a specific problem.
I like leetcoding, it has definitely improved my coding skills. Even if you have been writing code for many years, sometimes writing boilerplate code will not help you grow ability to write code creatively. I think doing leetcode type questions helps with that.
cool opinion, you're wrong
@@gackerman99 I am not.
@@Axl124124 you absolutely, categorically are
Its honestly such a fun way to get into a new language too. Never used x or y language, jump into some LC questions with it and enjoy.
Do you consider LC the best platform to focus your energy on, or do you also use others like HackerRank, Exercism, etc.? I've never used any of them TBH, but I think it'd help me do better in technical interviews, even after 12 years in industry.
That does make sense, however in this current job market they are expecting you complete the problem 100% Also now these companies use third party companies to do these technical interviews and ya if you don't complete 100% in the allocated time, you don't go through to the actual engineers with "the secret handshake". In my experience of only a month of job hunting, so far, and studying these leet code stuff, I am just really bad at these word problems but thankfully I know how to build a real project to keep my motivation going. Thank you for this bit of knowledge though, I am going to use it going forward for these interviews, cause sadly, they are all needing this 100% complete handshake.
It’s kinda how everyone I’ve met that uses Vim is a great programmer. Not because Vim makes them great, but because the kind of person that takes that much time to get good at it will probably put the same effort into everything else. Someone who’s good at coding challenges is likely going to put it work in every other skill
How about VS Code Vim like me. I love V motion but wont setup for my project.
Altough i felt like superior when editing config through ssh when looking down nano users
lol Fallacies all the way down...
@@patrickheney9201 why do you think so?
Because a portfolio of actual development job experience or contributions to a code base clearly doesn’t matter in *checks notes* a Software Engineering position.
Instead, let’s have you solve problems that you’ll never even tangentially interact with and that have been optimized for years?
I've been hiring software engineers for the past 10 years. I have a better method. Give them some dirty code and ask them to clean it up and make it more object oriented. This reflects 90% of the work they'll be doing day to day. If they can effectively do this they're hired.
Object oriented means it's dirty.
I hear this a lot that companies do this to see if you're good enough to do the job but being good at interviewing doesn't show you are good at the day to day work, or working in a team, or dealing with tight deadlines so isn't it still a risk either way? I do/did the handshake anyways because it is what it is at this point and I was hungry to change my situation but saying this handshake is necessary to know I have a good pick isn't really true right? I wonder if a probationary period, similar to an internship, is a better method. Idk a better way either, but I don't think the handshak is the best way to weed out the best candidates. I also wonder what other methods have been tried in the past🤔... I know some companies are trying different options now, which is really good, and I hope it can lead to positive findings.
Sounds like the editor doesn’t know the difference between a map, “four” and filter 😂
Been looking at leetcode as the MCAT, that helps with the “why do I have to do this”
In Germany the interviewsrevolve around real life use cases like API integration, data transformation, etc. No DSA time waster.
Imagine showing up to the hospital for an interview, and they say, " in order for me to know if you are really good doctor, please operate on this patient, in the next 20 minutes"😂😂😂😂
hearing that he expects that you’ve done 300 leetcode problems is so disheartening lmfao
For a faang job. Normal jobs you just do 50 easy ones, and cycle through a few interviews until you pass one.
If you don’t have the solution, at least talk through what you could do! Plan the system, even if it is only pseudocode or uses black boxes
Yes the secret handshake is stupid. Is like nepotism. I refuse to do leet code problems because they dont represent reality. I can waste 1 month learning all the answers and then after i get the job i am exhausted from all that learning.
A lazy interviewer will use leetcode problems.
A good one will find a few questions personalized for each candidate that shows how good they are.
When i need a team member that just have to do repetitive tidious work i might use leetcode. When i need problem solvers i run from that shit.
And if you have tidious repetitive work to do you might just as well get someone with problem solving skills to automate that process.
Its not telling the company how well you can code, its telling them how willing you are to jump through hoops.
I see two reasons:
- A short interview requires some kind of pass/fail task, even if it isn't particularly well correlated with the job.
- Nerds can't help but gameify everything.
Im a couple of years into my career and work w the NOAA as of recently. I've never used leet code more than a few times. I consider myself very capable and able to tackle any problems thrown at me. However, I I hesitate to apply to a FAANG company because I feel like I would bomb the leet code portion. I use so many tech stacks and frameworks and languages, It's hard to keep track of things like syntax and such, and it worries me. On the other hand, because of that, I'm not afraid to learn new languages and tackle challenging problems because I can recognize the similarities in all of these languages and I feel like it makes me a really strong candidate.
90% of the time, you are going to be hiring someone who hasn't done what you want them to do. So instead, companies need to figure out if the person can do what they want them to. So adaptability and ability to start to approach and solve problems is the thing they'll try to test. And that is not easy.
Since the questions can sometimes be randomly selected would it ever be okay to ask for another problem if they give you a problem statement and you don’t know that one? Just say “here’s my initial thought but getting into into this I am not quite sure about going down this path. Do you have any other questions prepared?”
Is that "map, *fold, and filter" - or "map, *for, and filter" - or something else?
300 questions aint getting you shit in this day and age.
Its not even the hoops, its the vibe
I'd rather know people understand writing loose coupled and clean code than solve some leetcode problems.
Nice ... So what are map four and Filter
you can take this knowledge and shove it. i have 20+ years in development and all i want to know during the interview, is whether i can reason with you. i would never hire this guy, cause he loves the sound of his voice too much.
What’s stopping me from just getting good at leetcode and not even knowing how to code really
That’s impossible
@@ci6516 there effectively math puzzles. You just then need to learn the syntax
@@SaceedAbulyou'd probably end up better in terms of fundamentals than the average beginner
Sorting algorithms is simply just working with numbers in a set that fit into an ordering different sorting algorithms might have different orderings.
Memorizing the syntax of how to code them is different from understanding what merge sort has to do with total ordering and how this can be abstracted in the first place.
LeetCode doesn't teach this.
My company gave me an assignment.
I had a week to build a PWA shopping app in accordance with a predefined design and integrate it with an existing API.
So while there is no good way of gauging how good someone is at coding in 50 minutes, who says the process needs to be confined to one interview?
No offence, but the only companies who've ever given me leet code interviews are shitty companies I've walked out of.
I feel that way about companies that hand out unpaid assignments. It's a red flag.
I hate take home assignments like yours. After spending an entire week they will simply reject without any feedback nor will give any monetary compensation for the time I put in. Its much better to do a "standardized" leetcode interview in your language of choice.
Like for example if you have never worked with java techstack, you won't even be considered for takehome styled interview, however with leetcode you can choose any language and give the interview regardless of if the company uses java or any other language/tech stack
I prefer problems that give the candidate multiple sets of data that has some relationship and then ask them to solve various problems by merging, joining and filtering.
The problem with leetcode questions is that they appear to be good if you're a bad interviewer, but are not actually good interview questions. So they have become a common trap bad interviewers/tech companies fall into, like something a junior interviewer would do while a senior would know better.
I've NEVER done Leetcode for an interview and never will. LC is like asking a mechanic about the mathematical equations for the physics for internal combustion.
EDIT: you know what I meant. Leetcode is NOT what the job entails. It is pretty much pointless. When was the last time you inverted a binary tree at work?
100% agree. Are we going to be only working on random questions everyday? such as: "Write an algorithm that figures out how many bananas are being eaten by 10 year olds who are watching TikToks."
... didn't think so.
not quite true
LC just doesn’t yield practical experience. At least the 5 year react dev that doesn’t know array functions does know how to use 3rd party libraries, and that’s a better skill most of the time at least unless you’re making the 3rd party library to begin with, i suppose.
@@nonamenolastname8501 they don't? That should be the interview! 😂
@@nonamenolastname8501you know not one mechanic I guess.
I've been coding for 20+ years. I've coded, led teams, architected solutions, hired teams, all of my projects have been successful... and I've never even heard of Leetcode.
I've seen professional students in a number of different fields, and they're the worst to work with. They think they know stuff, but can't actually do the job that's asked of them when they need to solve a real problem, and they lack the analytics skills to evaluate why one solution might be better than another. I feel like LC caters to people who study for rote memorization without the thinking skills needed for the real world. I want someone who understands the problem, someone who can provide a solution someone else might not have thought of... Not someone who gauged all their answers by what was in the back of someone else's book.
Sound like another person who didn’t go to uni who’s mad self taught is no longer an option and that boot camps are failing to produce results
It’s called DSA. Sorry ur not taught that at 15k bootcamp
@@ci6516 Lmao I already have my job. Why would I be jealous that people who can't hold my job have already moved on?
@@ci6516You have just ironically proved his point.
If you don't check basics, your managers while hire folks who don't have any.
Same for Upper Tier SOC and security analyst...
Have you memorized enough ports, attacks, ect...
Interviews are tough. Hard to determine aptitude.
But if you can't build Spotify type app BUT you can do a million Leetcode problems. Why are you being hired? I find knowing how to put all this stuff together into something working is a skill on it's own.
They want to hear me talk?
Over the online-with-no-intervewer-leet-code-challenge technical interview?
Damn, should have known that I had to record my voice and send them an email with the file afterward...
That's the secret!?!? 😂
map 4 and filter...this is adorable
Baaam… the knowledge hits me
ok this actually changed my mind about these questions. I always thought of them as just iq tests, essentially, except ones that are easy to cheat on and not indicative of much wrt how you'll be as an employee (except insofar as intelligence helps with everything). but testing for hard work and preparation kinda makes it seem more reasonable, especially as a way to counterbalance intelligence - you have to be either very smart, very hardworking, or somewhere along the function between the two that allows you to compensate with one for lack of the other.
I think leetcode is a great complementary skill to any programmers. it strengthen problem-solving skills and proficiency in DSA concepts like array, sets, maps.
I think those with good DSA skills can learn new tools pretty quickly. I learned a bit of React after 2 days and could already build something look somewhat good (but the code and folder structure behind is disgusting lol, lots of DSA-ing gave me the mind power to not be afraid of potentially typing hundreds lines of code)
why not demonstrate your ability to learn new tools quickly by having multiple projects that use different tech stacks that you can talk about in-depth?
I would rather start my own business. If you can code something that solves real problems THEN YOU DON'T NEED A FREAKING JOB>
I never ask leetcode questions. I asked scenario based questions based on what they have written on their resume. My favourite one is to write a class with properties that can store data from an xml file. Simple question that test both knowledge and problem solving.
It's actually illegal to give an iq test, but a coding question is the same thing.
Never used a binary tree search algorithm in my 6 years from frontend dev
Leetcode also scales from the company's POV. If you have 50 people interviewing 200 engineers, for 10 positions, striving for a consistent grading process for all 200 interviews is really important. Way more consistent than "tell me about what you've built" at least.
“I’ve built nothing but LC problems, filling in those functions are my apps, I’ve never used an external library in my life, I always recreate the wheel”, yeah… that’s real trust inspiring…
The only consistent thing you get out of doing that is you can happily throw away a lot of those applications. That's what Leetcode is REALLY doing, giving people a convenient way to narrow shit down.
But at least they can invert a binary tree instead of center a div in CSS
How does “I copied RUclips tutorials projects prove anything
I think if you are able to gring 300 lc problems. You can learn anything about the job. Because grinding that is extemly boring so at a real job it wont be that bad
I love leet code problems more than actual problems 😂
Our local big techs say "you have to be able to code without bugs so that you spend less time debugging when you write our production code". I call BS.
I've chosen not to hire people who have all the answers without thinking about it before, drag that handshake on to make it seem like you're solving it in real time vs reiterating the answer
i respect your code editor background image
If you can afford it then take home interview projects are the best way to guage their knowledge.
With take home projects you actually go beyond just solving random abstract questions which are usually found on leet code and put the interviewee at the test for the things they're actually going to work for years if they get selected.
Some important things which are just hard to guage from a pure LC interview - Code quality, version control, testing, practical problem solving.
I don’t think they need to be take home challenges. You can present a candidate with a simpler issue that can be solved, or at least have them explain how they would go about solving it. The challenge can be representative of what you do in the job, you get a notion of how the candidate thinks and communicates, that’s a lot more valuable than memorizing some algorithms for LC. Plus it respects their time I wouldn’t expect a candidate to dedicate multiple hours of their own time to solve a problem when chances are they’re interviewing for other jobs as well
I disagree with this one. Much better to show them some of your system's code, and have them describe what it might be doing and why it might be doing it that way. Forget the leetCode test garbage.
@redfordkobayashi6936 Even leetcode easy requires things that many senior devs haven't ran into. If you ask someone to do an Inorder Traversal of a binary tree, the chances that they ran into a single problem that required them to write a binary tree in their day to day life programming professionally is pretty low.
And for good reason. A binary tree is probably not the best solution for a lot of their day to day problems.
So they're going to be stuck on what "Inorder traversal" even means for a bit.
Of course, this all depends on the job you're applying for. If you're applying for a job at a company in which such data structures are common place then yeah you should be expected to understand them.
But even then, just showing your actual code to the person and having them describe what it's doing? Accomplishes both.
If such data structures *aren't* needed. Then your sample won't have them.
If such data structures *are* needed. Then your sample will probably have them.
Is this possibly more difficult? Yes. But it's solving the core problem of the question being relevant to the actual job.
The problem is a very high number of poor candidates who cannot even solve a simple problem. Use LC or similar for screening and then pair programming for the actual interview.
@@zlamanit In the context of a 30 minute window, what exactly is the "problem?" Can you define it any better?
The reality is that hard work is what makes most people good at a job. If you've grinded 300 leetcode problems, you have, at the bare minimum, established that you are have some work ethic to you.
People don't like it, but showing up and giving consistent effort over time puts you above a significant % of other candidates.
Slogging four years at university isnt enough?
Well 90% of the coders at Google suck so how do you explain that? They have one product they designed 25 years ago.
If you can spend several months alone to build a project from the ground up then for me that is much more impressive that any grind you can do on leetcode.
for the love of god can we figure out a certification system, like the FE and PE engineering exams?
what the hell is a map four
I'm still on map 3
map, for each(shortened to for) and filter, the subtitles are probably automated
maps are just arrays that dont confine their indexes to integers only, you can use strings and such to store variables, think
[ 'price' => 123, 'description' => 'zxcxc', 'sometingElse' => [ other map here ] ]
instead of
[123, 'zczcx', [ other array here ] ]
I usually ghost companies when they try and give me tests or assignments. Maybe if I was unemployed I'd have time for that but I'd rather go outside after work and do non work related shit.
Yeah, my first ever job didn't involve a technical portion. I coded a fair amount in that position, too.
The thing I don't get about this, is that these companies are highering programmers who will work for them for months or maybe even years. Doesn't that mean it would be worth whatever time it needs to interview a person in a more representative setting?
This could mean asking them about how they handled past projects and maybe give them a more complex problem with the required time to solve it like one would on the actual job.
The amount of time you waste with programmers that might have been better on the leet code tests than the actual job or even worse the talented people you loose to other companies should be a major factor when coming up with an interview process.
A company wanting me to jump through hoops to see if im "good enough" is a red flag to me, and im not interested in working at a place like that.