Finally, a sane mind has spoken. People don't seem to realize how time consuming these pieces of "homework" are and how subjective they are going to be graded, especially given that companies don't provide feedback. Next year, candidates will be begging for the return of Leetcode.
@noobgam6331 Not really. I don't do low level code for a living but I do some fairly advanced microcontroller programming in my spare time and it would be very rare that I'd implement an algorithm from scratch. Even the linux kernel has built in routines for linked lists and such things so even kernel developers would rarely if ever implement algos from scratch. I think you're missing the point. It's not a low-level vs. high-level thing. It's about memorizing a bunch of algorithms that you'll probably never use vs. developing the skills necessary to thrive in the IT industry by diving into big, mature codebases and being able to get up to speed and understand the major components and map it all out in your head. Any yes, fixing bugs.
@@noobgam6331probably you never worked on low-level programming. These things are pretty standard , you don’t reinvent. For example: you don’t implement a sorting algorithm, you just implement the best that is already present
@@nakatash1977 no. Some people actually /have to/ write the libraries. Recognizing a lot of algorithms helps a lot. When I encountered a problem in a high-level project that just uses a 2-pointer technique (we needed to build something over a sliding window timeline for SVOD licences), none of my colleagues knew such a thing even existed and it would've been a big problem. Which I'm pretty sure they would solve via map-reduce cluster instead and consume a lot of resources for no reason if I weren't there. To this day I believe that algorithmical foundation is the core thing that reflects your computer science knowledge (not the knowledge of system design/how to work with an actual codebase). So if the process would be to completely eliminate algorithmical challenges from inteview process - I'd be hard against that. If you add the bug search ON TOP of that - it's a good addition. P.S. I do work at FAANG companies from the age of 18. Never applied to smaller ones, and a lot of the project I worked on involved leetcode-hard level problems. (Actually after transitioning to a few years ago to AWS I haven't used a single algorithm at all, but that's a different story)
i do think while linked lists arent that used (theyre not that good data structs anyways), graph and tree algorithms are, and ive personally done a few from scratch tree walks based on the problem domain because of that i dont think leetcode algos are fully "toy and useless to learn because no one does them from scratch"
It's great that they're moving in this direction, but I hope long-winded take-home projects don't become the norm. I think it's ridiculous to have to spend 4 hours of your own time building/completing some project only to potentially not even pass the resume screen.
I think it's also ridiculous to grind 500 leetcode problems and still don't pass the resume screens. It is just this way, you have to take a bet whether the time spent on anything is going to be fruitful or not.
4 hours I can live with. I've been given take homes that if it were handed out in sprint planning would have been broken into 2-3 tickets and would likely be estimated at a week. I'm not going to work for more than an afternoon for free that's for sure
While I’m not a big fan of take-home tasks, I prefer them over online assessments that track your every move, 100 hours of LeetCode, or coding exercises where you’re expected to recall everything from memory. These approaches don’t reflect the real-world job environment. Experienced developers, data scientists, and other professionals will tell you that effective coding is a combination of innate knowledge, the ability to read and debug code, and skillful use of external resources-whether it’s Stack Overflow, RUclips tutorials, forums, or large language models. This blend of knowledge and resourcefulness is far more representative of one’s ability to perform in a role. While take-home tasks have their flaws, they can contribute to a portfolio and provide an opportunity to demonstrate skills in a more meaningful and realistic context. Framed appropriately, they can even offer long-term benefits for candidates.
Take home tasks take time, which you can't do do dozens of them, meaning you'll apply to less companies and choose more carefully to apply to those where you're a good fit and you're actually interested in the job. I think it's a great filtering tool
@@exoneuromancer1672 I think in reality this is not going to happen. You are assuming applying to less job will still get you the same amount of interviews, which imo is unlikely. People who are desperate will still be spamming jobs.
Be careful what you wish for. This will inevitably turn into something ridiculous. Top companies need a way to filter down thousands of candidates so they can maintain their 2% acceptance rate. This will be adapted to make sure very few people can succeed, even if the end result is as ridiculous as people see leetcode.
This just filters out a lot of people that pass simply because they memorised a solution or did it a thousand times without questioning it. I hate recursion with a passion. Leetcoders love it and they don't seem to understand why it's so bad. Most of them are unaware of basic architectures and they tend to skip and rush ahead. I know because I've done these interviews tons of times and I'm sick and tired of cleaning up messy leetcoder code.
@ I get what you’re saying. I’m not really disagreeing with that. What I’m saying is that it’ll likely come full circle and those same people will have success with the next system. How do these companies take thousands of applications and hire 2% or less of those applicants without having ridiculous aspects to their interview process? They’ll probably have us memorizing design patterns and have you fix a bug within 30 minutes in a 1-million line code base with tons of domain-specific knowledge within it.
@@chudchadanstud ? Recursion is not bad at all. Recursion is just a way of using the stack of the thread rather than allocating and using your own. If your take is "recursion is bad" and not "recursion is a tool to solve problems with pros and cons" YOU are the problem here.
@@josephp.3341 No, recursion is inherently risky precisely because you're relying on the call stack. The stack is limited in size. Each recursive call involves pushing a new stack frame, which can lead to unnecessary memory allocations and is particularly dangerous when the recursion depth grows, as it can cause a stack overflow. While recursion can be elegant and concise for certain problems, it often trades off scalability and performance. This is why I don't like Leetcoders. They don't understand nuance. It's no different from fizzbuzz.
@@josephp.3341 Don't know why RUclips deleted my comment. No, recursion is inherently risky precisely because you're relying on the call stack. The stack is limited in size. Each recursive call involves pushing a new stack frame, which can lead to unnecessary memory allocations and is particularly dangerous when the recursion depth grows, as it can cause a stack overflow. While recursion can be elegant and concise for certain problems, it often trades off scalability and performance.
I hate take homes. They always seem to be a whole feature that would count as deliverable with concrete value. Sometimes I'm wondering whether the job is for real or whether I'm just working for free. I won't do anything I would estimate as more than a couple of hours work now and much prefer if I am pair programming with one of their engineers or leads, then at least I know they are taking the test as seriously as they are asking me too!
Man, I'm a fresh graduate and I've had to take a few take homes. It's so bad when you spend hours on an exam and you get rejected with 0 feedback, you dont even know where you could improve to finally get a job.
The technical interview I had for my current job was a 3-hour take-home and use any resources I wanted to as long as I could discuss the code and its tradeoffs at the end. I definitely feel that it more correctly gauged my ability to write clean and maintainable code than a 45 minute leetcode question. Ultimately on the job, that’s a more valuable skill than being able to shit out Dijkstra imho
Well thank god. As someone who has worked as a software development consultant at a few companies, 80% of my work involved things that were less algorithmically difficult than a leetcode medium. The difficulty for me came most times in contributing to pre-existing code which involves combing through 1000s of lines of code/ multiple microservices, figuring out where changes need to be made, fully understanding what previous people have done that will then allow you to integrate new code without breaking anything.
exactly, and they only want to filter people based on how good they are at problem solving and they train them after they hire them, because why not? they have billions of dollars.
While this is good way to test. But the good thing about Leetcode style problem is to come up with new problem sets easily. While on the other hand Hackerrank will have to make a code base again and again so that candidates cannot copy things from web. I may also be wrong but Hackerrank is primarily used for screening and major organisation will still not go in this direction. They just want to capture the market that uses assignment instead of Leetcode problem. But yeah very hard to keep coming up with new codebases😢
Don't think leetcode does that. It just makes you memorise something you can google. It's like hiring someone because they memorised a slightly more complex fizzbuzz. Being able to review a piece of code, ask questions, critique it and come up with a better solution is way more valuable. You can't memorise that.
Great to see this transition in interview styles. The few companies I've interviewed that gave a repo problem like this have been the best experiences I've had as a candidate. btw: PRD = Product Requirement Document. Basically it's a spec sheet from a product manager of the big picture idea and business requirements for a feature. Which then is expanded on by engineering into a TRD Technical Requirement Document (aka a ticket) for an engineer to work on in their sprint.
Cant think of any obvious downsides to this, other than the potentially awkward transition out of leetcode style hiring. I will be interested to test it out. Congrats on 100k by the way.
My concern with this strategy is that it will take a long time as a candidate to go through all of those steps if there is to be significant enough depth to really test a candidate, and the results are more arbitrary. With leetcode style problems they aren’t as practical, but they are more binary in terms of pass/fail.
Am I the only one who thinks LC skills actually help at work? I have only 2 YOE and have worked on CRUD apps on run of the mill companies but I’ve come across a couple of situations where I needed to implement some sort of LC-style algorithm
Yes but only from time to time the vast majority of work is much closer to a take home to implement some feature, business knowledge abstraction or just plain old bug fixing rather than explicit data structure/alg implementation into a super specific technical question. It does help knowing your space and time complexity though I've seen some monstrosities of triple nested for loops or recursion that could be solved extremely simply with simple data structures.
Wouldn't say so. Leetcode fundamentally gets you good at programming and solving problems universally. Not every company is focused on web development. Leetcode is great man plus makes the other things look easier to grasp and understand.
Leetcode is actually better cause it’s way less about knowing some specific framework. It’s more like basic skills that don’t change much over time. Big tech companies are more into hiring engineers in first place, not frameworkers
What makes you think a take-home assignment or interview would require you to use a framework? Let’s see you build something with your solution to two sum lmao.
@@josephp.3341 yeah, and leetcode, plus the technologies a candidate has worked with - that’s usually how it goes. i meant software engineer, not a software architect, though, it was intentional sophistication to describe someone who has deep knowledge in the field
@@Bad4Good2 bro, of course you’re gonna use some framework or library for almost every feature, unless you wanna waste a ton of time and clients' money on a custom solution. plus, in the video, there was an interview with tasks on the mern stack. about leetcode, i just meant that frameworks can change, use different stuff under the hood, and you wouldn’t always know, but two sum will always be two sum. and yeah, i think understanding some of these problems might not help much on the job, but it’ll definitely help you grow as a computer guy.
this is great news, I recently just bombed an interview because I couldn't solve a leetcode hard problem. I spent more time building projects than grinding leetcode.
The solution to LeetCode monkey style interviews is simple, just refuse to have those. Mid difficulty algo questions made sense 20 years ago, now everyone knows about them and with tough job market it become a mindless race to the bottom.
LeetCode isn’t about IQ, it’s mostly about pattern memorization. It’s like knowing the recipe for every kind of dish. It’s all about recognizing the menu, not inventing new ones. It teaches you how to solve a specific type of problem quickly, but that’s not necessarily the skill businesses need. They want you to solve real-world problems with the right abstractions and tools, not reinvent the wheel.
@@4sxS307cAWHow do you think you learn, by understanding everything from scratch? It's not all about memorizing, you see different styles of problems in each interview so you have to understand it.
this is overcorrecting or swinging extremely to the other end. tests should be generic capturing the understanding of CS foundations and concepts not how many commands you know of some tool. plus take homes need to be generic so companies also value candidates time... i don't know may be I guess we are asking too much of companies 😅
I have been told by my peers that I am the best cpp dev and I was never told to do a radix sort lol😂. Nonetheless some data structures, graphs questions are useful in the sense that if you cannot solve them in time it would be problematic. I think many questions in hacker rank can be solved if you do work a lot in optimisation problems.
I'm working as a software engineer for the last 5 years, knowing computer basics like ds and Algo, is, networking, dbms is essential but not to a point where we are constantly solving stupid recursive problems and dynamic programming... This is actually not the way to hire a candidate
Implement Vector’s push_back method is a way better DSA test than “Sam can rob from every house, but can’t rob from adjacent houses, maximize the amount of money Sam can rob from the neighborhood.” One tests your understanding of a container that is used ubiquitously in day-to-day programming (and can be reasoned through even if you’ve never done a similar problem). The other tests your ability to have enough free time to grind out similar problems that are not applicable to real world programming.
@CodingJesus completely agree ... Software engineers are given brain teasers now a days.... I work in a system framework team I see crazy bugs in mono, low level platform abstraction bugs, memory issues.... It has nothing to do with these brain teasers.
@@CodingJesus Implementing a Vector's push_back is simply an "insert a node to the end of a Linked List". There's a reason why companies ask harder questions such as "rob from adjacent houses". The bar is higher due to competition.
@@bufdud4 no. a vector is a dynamic array and push_back requires reallocating and copying memory -- usually in an exponentially growing fashion. This type of question imo better probes peoples knowledge.
@@CodingJesus Hard agree. If I had to give an interview I'd do this: Only a single programming question: something involving maps/lists/vectors. Then I'd do a code review of one of their projects, where I'd ask them to justify/weigh pros and cons of any decisions (architectural/code) in the project. Then I'd end with a systems design question that takes the form of a conversation. #1 checks you know how to code in the language you're getting the job in #2 checks you have the ability solve harder problems and that you can communicate those ideas to other developers #3 checks you have abstract problem solving skills and you know how to apply them to the real world Somehow hiring people can't figure this out
how come a doctors are not asked to take home quizzes on how to solve brain surgery questions before joining a hospital and they walk out with 400k a year
Just admit it. The industry is cooked. Think about it from a business perspective: if you had millions of contracts, would you make it easier or harder to hire people? EASIER because at the end of the day, you need one million people (for example) to complete those contracts. It becomes a numbers game. You dont give a crap if someone can answer technical questions as long as they can do the job and fulfill the contracts. The problem is, companies dont have millions of contracts. They maybe have 3 big ones and that's it. TL;DR Companies don't have enough jobs to hand out.
LOL this is the dumbest shit. HackerRank's take on Google's whiteboard interviews must be "once keyboards are invented they will revolutionize interviews".
This debate will never end… Leetcode has never been about testing software engineering skills. It has always been about testing IQ and it’s pretty good at that. It’s also a good way to test how much a candidate has “grinded” for a position. Now are those 2 good measures of how well a candidate will perform? Maybe, maybe not, but what is certain is that there is no way to test that out in a simple interview process of 3-8 meetings. Their new “fix” isn’t a magical solution either.
IQ? Are you joking? Solving leetcode has nothing to do with IQ, it is all about memorizing the answers, thats it. Same as entrance exams etc. It favors people who memorize, not the smart people.
@@2A9D8F And who can learn more problems faster / understand and memorize harder problems? People with a higher IQ... If it takes you 20 minutes to understand an easy you're gonna cover less problems than the guy who covers it in 30 seconds Same with entrance exams: those who succeed are either smart or have grinded hard asf. I'm fine with hiring either.
lol tell that to ppl who got lucky that they got asked a similar leetcode they practiced the day before. Also meta just makes you memorize the top leetcode Meta questions. Lets stop pretending that devoting your spare time to grind leetcode and having no life means you have a high IQ.
Havn't watched the video, but I grew to really enjoy leetcode and like it more as an interview process than an interview where I have to implement some feature or fix bugs on the spot. To me, that is much harder because its harder to directly prepare for that process, while with leetcode there is a direct roadmap and process to prepare for interviews and with enough time/hard work it is possible to master interviews.
People be hating leetcode then join a workplace and write shitty code. Just because your code works, doesn’t mean it’s correct. And if you can’t thing when your problem has a very very limited scope, there is no way you are going to be comfortable working in a huge codebase. Stop fooling yourself, there is no way around leetcode. All these videos are a clickbait to attract losers who think they are too good to practice leetcode
LeetCode can definitely help with problem-solving skills, but it's not the only thing that makes a good developer. Real-world coding is about writing maintainable, efficient code that solves actual business problems. Just because someone doesn't excel at LeetCode doesn't mean they can't write good code. Also, working in a large codebase requires more than just algorithmic thinking - it’s about collaboration, design patterns, and being pragmatic with your solutions. LeetCode is one tool, but it's not the end-all be-all.
Leetcode has a very limited scope but any somewhat capable dev can brush up on data structures and algs and grind out neetcode 150 in a few months. Doesn't mean they can get a system running that actually adds value to a business.
@@4sxS307cAW True af. I have a decently paid job in a bank and I've literally never done a leetcode problem except in the actual interview to get my job. On that actual job 99% of my effort is spent on communication, systems design, architecture, planning, product, database, operating systems, concurrency, testing, logging, deployment or literally any else that is not related to how to find whatever in a bipartite graph.
Bruh what's with the hate for leetcode questions. I get that memorizing questions is useless for an SE job. But if that's what you feel is the problem. Just make the problems harder. Solving hard competitive programming style problems is a really good test of someone's problem solving skills. And if someone has good problems solving skills, (and really good implementation skills thanks to coding up algorithms you come up with to solve hard problems). I can guarantee you that person will learn way faster than a candidate who only does normal development, and fix bugs that they encounter.
Most companies problems aren't really about how to structure data and most companies have little to no interest in improving performance at a faster rate than what they currently have. But almost all companies have UX issues in their design, and features that while seemingly simple, actually affect more of the product than we realize. Some companies even have extremely basic UX, design issues and missing features and it's worth spending the time in the interview verifying if they will be able to contribute to it for sure rather than testing how they can solve a problem that is essentially a single function that runs in the context of a single file.
Why the hate in leetcode? It's great fun and challenging. Leetcode shouldn't be used to memeorize a problem but understand how problems work and common approaches to them and it becomes a natural way to break down complex problems and is universal knowledge not just web dev.
Hackerrank concluded you don't need leetcode style problems, you need hackerrank style problems.
Finally, a sane mind has spoken. People don't seem to realize how time consuming these pieces of "homework" are and how subjective they are going to be graded, especially given that companies don't provide feedback. Next year, candidates will be begging for the return of Leetcode.
Hii
I've been working in software for 10 years and have never professionally wrote a linked list. But I have fixed many bugs.
That just means that you haven't worked on any low level code whatsover.
That happens in many industries but not all.
@noobgam6331 Not really. I don't do low level code for a living but I do some fairly advanced microcontroller programming in my spare time and it would be very rare that I'd implement an algorithm from scratch. Even the linux kernel has built in routines for linked lists and such things so even kernel developers would rarely if ever implement algos from scratch.
I think you're missing the point.
It's not a low-level vs. high-level thing. It's about memorizing a bunch of algorithms that you'll probably never use vs. developing the skills necessary to thrive in the IT industry by diving into big, mature codebases and being able to get up to speed and understand the major components and map it all out in your head. Any yes, fixing bugs.
@@noobgam6331probably you never worked on low-level programming. These things are pretty standard , you don’t reinvent.
For example: you don’t implement a sorting algorithm, you just implement the best that is already present
@@nakatash1977 no. Some people actually /have to/ write the libraries. Recognizing a lot of algorithms helps a lot. When I encountered a problem in a high-level project that just uses a 2-pointer technique (we needed to build something over a sliding window timeline for SVOD licences), none of my colleagues knew such a thing even existed and it would've been a big problem. Which I'm pretty sure they would solve via map-reduce cluster instead and consume a lot of resources for no reason if I weren't there.
To this day I believe that algorithmical foundation is the core thing that reflects your computer science knowledge (not the knowledge of system design/how to work with an actual codebase). So if the process would be to completely eliminate algorithmical challenges from inteview process - I'd be hard against that. If you add the bug search ON TOP of that - it's a good addition.
P.S. I do work at FAANG companies from the age of 18. Never applied to smaller ones, and a lot of the project I worked on involved leetcode-hard level problems. (Actually after transitioning to a few years ago to AWS I haven't used a single algorithm at all, but that's a different story)
i do think while linked lists arent that used (theyre not that good data structs anyways),
graph and tree algorithms are, and ive personally done a few from scratch tree walks based on the problem domain
because of that i dont think leetcode algos are fully "toy and useless to learn because no one does them from scratch"
It's great that they're moving in this direction, but I hope long-winded take-home projects don't become the norm. I think it's ridiculous to have to spend 4 hours of your own time building/completing some project only to potentially not even pass the resume screen.
I think it's also ridiculous to grind 500 leetcode problems and still don't pass the resume screens. It is just this way, you have to take a bet whether the time spent on anything is going to be fruitful or not.
4 hours I can live with. I've been given take homes that if it were handed out in sprint planning would have been broken into 2-3 tickets and would likely be estimated at a week. I'm not going to work for more than an afternoon for free that's for sure
While I’m not a big fan of take-home tasks, I prefer them over online assessments that track your every move, 100 hours of LeetCode, or coding exercises where you’re expected to recall everything from memory. These approaches don’t reflect the real-world job environment.
Experienced developers, data scientists, and other professionals will tell you that effective coding is a combination of innate knowledge, the ability to read and debug code, and skillful use of external resources-whether it’s Stack Overflow, RUclips tutorials, forums, or large language models. This blend of knowledge and resourcefulness is far more representative of one’s ability to perform in a role.
While take-home tasks have their flaws, they can contribute to a portfolio and provide an opportunity to demonstrate skills in a more meaningful and realistic context. Framed appropriately, they can even offer long-term benefits for candidates.
Take home tasks take time, which you can't do do dozens of them, meaning you'll apply to less companies and choose more carefully to apply to those where you're a good fit and you're actually interested in the job. I think it's a great filtering tool
@@exoneuromancer1672 I think in reality this is not going to happen. You are assuming applying to less job will still get you the same amount of interviews, which imo is unlikely. People who are desperate will still be spamming jobs.
Be careful what you wish for. This will inevitably turn into something ridiculous. Top companies need a way to filter down thousands of candidates so they can maintain their 2% acceptance rate. This will be adapted to make sure very few people can succeed, even if the end result is as ridiculous as people see leetcode.
This just filters out a lot of people that pass simply because they memorised a solution or did it a thousand times without questioning it.
I hate recursion with a passion. Leetcoders love it and they don't seem to understand why it's so bad. Most of them are unaware of basic architectures and they tend to skip and rush ahead. I know because I've done these interviews tons of times and I'm sick and tired of cleaning up messy leetcoder code.
@ I get what you’re saying. I’m not really disagreeing with that. What I’m saying is that it’ll likely come full circle and those same people will have success with the next system. How do these companies take thousands of applications and hire 2% or less of those applicants without having ridiculous aspects to their interview process? They’ll probably have us memorizing design patterns and have you fix a bug within 30 minutes in a 1-million line code base with tons of domain-specific knowledge within it.
@@chudchadanstud ? Recursion is not bad at all. Recursion is just a way of using the stack of the thread rather than allocating and using your own.
If your take is "recursion is bad" and not "recursion is a tool to solve problems with pros and cons" YOU are the problem here.
@@josephp.3341
No, recursion is inherently risky precisely because you're relying on the call stack. The stack is limited in size.
Each recursive call involves pushing a new stack frame, which can lead to unnecessary memory allocations and is particularly dangerous when the recursion depth grows, as it can cause a stack overflow.
While recursion can be elegant and concise for certain problems, it often trades off scalability and performance.
This is why I don't like Leetcoders. They don't understand nuance. It's no different from fizzbuzz.
@@josephp.3341
Don't know why RUclips deleted my comment.
No, recursion is inherently risky precisely because you're relying on the call stack. The stack is limited in size.
Each recursive call involves pushing a new stack frame, which can lead to unnecessary memory allocations and is particularly dangerous when the recursion depth grows, as it can cause a stack overflow.
While recursion can be elegant and concise for certain problems, it often trades off scalability and performance.
I hate take homes. They always seem to be a whole feature that would count as deliverable with concrete value. Sometimes I'm wondering whether the job is for real or whether I'm just working for free. I won't do anything I would estimate as more than a couple of hours work now and much prefer if I am pair programming with one of their engineers or leads, then at least I know they are taking the test as seriously as they are asking me too!
Valid
Man, I'm a fresh graduate and I've had to take a few take homes. It's so bad when you spend hours on an exam and you get rejected with 0 feedback, you dont even know where you could improve to finally get a job.
The technical interview I had for my current job was a 3-hour take-home and use any resources I wanted to as long as I could discuss the code and its tradeoffs at the end. I definitely feel that it more correctly gauged my ability to write clean and maintainable code than a 45 minute leetcode question. Ultimately on the job, that’s a more valuable skill than being able to shit out Dijkstra imho
I always preferred take-home tasks. 👍
Knowing how quick oranges rot is gonna be lost to the sands of time
Are you by any chance talking 'bout rotten oranges leetcode medium problem! 😅
@GoogleUser-uv9bo that is the bit yes haha
@GoogleUser-uv9bo Haha yes that is the bit
Well thank god. As someone who has worked as a software development consultant at a few companies, 80% of my work involved things that were less algorithmically difficult than a leetcode medium. The difficulty for me came most times in contributing to pre-existing code which involves combing through 1000s of lines of code/ multiple microservices, figuring out where changes need to be made, fully understanding what previous people have done that will then allow you to integrate new code without breaking anything.
Isn’t the reason why big companies don’t like to do this is because they often have their own custom stacks?
exactly, and they only want to filter people based on how good they are at problem solving and they train them after they hire them, because why not? they have billions of dollars.
While this is good way to test. But the good thing about Leetcode style problem is to come up with new problem sets easily. While on the other hand Hackerrank will have to make a code base again and again so that candidates cannot copy things from web. I may also be wrong but Hackerrank is primarily used for screening and major organisation will still not go in this direction. They just want to capture the market that uses assignment instead of Leetcode problem. But yeah very hard to keep coming up with new codebases😢
Don't think leetcode does that. It just makes you memorise something you can google. It's like hiring someone because they memorised a slightly more complex fizzbuzz.
Being able to review a piece of code, ask questions, critique it and come up with a better solution is way more valuable. You can't memorise that.
Great to see this transition in interview styles. The few companies I've interviewed that gave a repo problem like this have been the best experiences I've had as a candidate.
btw: PRD = Product Requirement Document. Basically it's a spec sheet from a product manager of the big picture idea and business requirements for a feature. Which then is expanded on by engineering into a TRD Technical Requirement Document (aka a ticket) for an engineer to work on in their sprint.
Corporations discover common sense
Cant think of any obvious downsides to this, other than the potentially awkward transition out of leetcode style hiring. I will be interested to test it out. Congrats on 100k by the way.
id say its hard to do this in the usual 1 or 2 hours, I hope long takehomes dont become the norm
My concern with this strategy is that it will take a long time as a candidate to go through all of those steps if there is to be significant enough depth to really test a candidate, and the results are more arbitrary. With leetcode style problems they aren’t as practical, but they are more binary in terms of pass/fail.
Hey man!! Congrats on 100k subs. love the content, please keep it up :-)
finally they've changed. I've failed mostly of these tests, and i've seen horrible code from ppl that passed them
Interesting to see those changes, great video. Also congratz on 100k subs!
PRD" stands for "Product Requirements Document
Thanks
Where can I find the pdf you are reading from? Thanks.
The sooner this rolls out and is used everywhere the better. The vast majority of engineering jobs are just some latest variation on doing CRUD
So it means people will go away from cp/icpc coz there wont be a purpose to it after i suppose😢😢
Is there any website or software where you can develop/backtest quantative trading strategies (maybe for free😬)?
TIL: PRD stands for Product Requirements Document
Feels like they are testing your git knowledge
That's extremely useful though. Developers with poor GIT skills are near useless.
Am I the only one who thinks LC skills actually help at work? I have only 2 YOE and have worked on CRUD apps on run of the mill companies but I’ve come across a couple of situations where I needed to implement some sort of LC-style algorithm
Some jobs still require that but majority don’t since they already implemented that or give seniors the duty to do that.
Yes but only from time to time the vast majority of work is much closer to a take home to implement some feature, business knowledge abstraction or just plain old bug fixing rather than explicit data structure/alg implementation into a super specific technical question. It does help knowing your space and time complexity though I've seen some monstrosities of triple nested for loops or recursion that could be solved extremely simply with simple data structures.
You were almost certainly unaware of a lib that will do it for you
this is good because i can do leetcode now without the pressure of having to get good enough at it for interviews
So is my leetcode grinding going to be a waste 😢
Wouldn't say so. Leetcode fundamentally gets you good at programming and solving problems universally. Not every company is focused on web development. Leetcode is great man plus makes the other things look easier to grasp and understand.
Leetcode is great when you're new. Past a certain point, learn to be an architect.
I don't think a company like Google would ever abandon algo questions. Keep grinding
Leetcode is actually better cause it’s way less about knowing some specific framework. It’s more like basic skills that don’t change much over time. Big tech companies are more into hiring engineers in first place, not frameworkers
What makes you think a take-home assignment or interview would require you to use a framework? Let’s see you build something with your solution to two sum lmao.
If you want an engineer you should ask systems design questions.
@@josephp.3341 yeah, and leetcode, plus the technologies a candidate has worked with - that’s usually how it goes. i meant software engineer, not a software architect, though, it was intentional sophistication to describe someone who has deep knowledge in the field
@@Bad4Good2 bro, of course you’re gonna use some framework or library for almost every feature, unless you wanna waste a ton of time and clients' money on a custom solution. plus, in the video, there was an interview with tasks on the mern stack. about leetcode, i just meant that frameworks can change, use different stuff under the hood, and you wouldn’t always know, but two sum will always be two sum. and yeah, i think understanding some of these problems might not help much on the job, but it’ll definitely help you grow as a computer guy.
this is great news, I recently just bombed an interview because I couldn't solve a leetcode hard problem. I spent more time building projects than grinding leetcode.
PRD == Product Requirements Document. Often issues by project managers to the development team when kicking off a new project.
The solution to LeetCode monkey style interviews is simple, just refuse to have those.
Mid difficulty algo questions made sense 20 years ago, now everyone knows about them and with tough job market it become a mindless race to the bottom.
Ridiculous that the interview questions are often 10x more difficult than anything most will ever do on the job.
Where is the link to the 8 page article?
wondering if codesignal will adapt
This is great approach. (i suck at leetcode and have zero iq)
Me too
LeetCode isn’t about IQ, it’s mostly about pattern memorization. It’s like knowing the recipe for every kind of dish. It’s all about recognizing the menu, not inventing new ones. It teaches you how to solve a specific type of problem quickly, but that’s not necessarily the skill businesses need. They want you to solve real-world problems with the right abstractions and tools, not reinvent the wheel.
@@4sxS307cAWHow do you think you learn, by understanding everything from scratch? It's not all about memorizing, you see different styles of problems in each interview so you have to understand it.
when information is freeflowing so much nowadays, leetcode tests if a candidate has been drilling a lot of problems.
He is the Messiah and finally companies have realized the truth of His message.
Is leetcode the dumbest thing we could ever do? No? Let's add AI. *chef's kiss*
lmao want to change the way they do things while me and countless others are in the middle/start of our grind... funny
this is overcorrecting or swinging extremely to the other end. tests should be generic capturing the understanding of CS foundations and concepts not how many commands you know of some tool. plus take homes need to be generic so companies also value candidates time... i don't know may be I guess we are asking too much of companies 😅
I have been told by my peers that I am the best cpp dev and I was never told to do a radix sort lol😂. Nonetheless some data structures, graphs questions are useful in the sense that if you cannot solve them in time it would be problematic. I think many questions in hacker rank can be solved if you do work a lot in optimisation problems.
100K let's go mewing Jesus!
I'm working as a software engineer for the last 5 years, knowing computer basics like ds and Algo, is, networking, dbms is essential but not to a point where we are constantly solving stupid recursive problems and dynamic programming... This is actually not the way to hire a candidate
Implement Vector’s push_back method is a way better DSA test than “Sam can rob from every house, but can’t rob from adjacent houses, maximize the amount of money Sam can rob from the neighborhood.” One tests your understanding of a container that is used ubiquitously in day-to-day programming (and can be reasoned through even if you’ve never done a similar problem). The other tests your ability to have enough free time to grind out similar problems that are not applicable to real world programming.
@CodingJesus completely agree ... Software engineers are given brain teasers now a days.... I work in a system framework team I see crazy bugs in mono, low level platform abstraction bugs, memory issues.... It has nothing to do with these brain teasers.
@@CodingJesus Implementing a Vector's push_back is simply an "insert a node to the end of a Linked List". There's a reason why companies ask harder questions such as "rob from adjacent houses". The bar is higher due to competition.
@@bufdud4 no. a vector is a dynamic array and push_back requires reallocating and copying memory -- usually in an exponentially growing fashion. This type of question imo better probes peoples knowledge.
@@CodingJesus Hard agree. If I had to give an interview I'd do this:
Only a single programming question: something involving maps/lists/vectors.
Then I'd do a code review of one of their projects, where I'd ask them to justify/weigh pros and cons of any decisions (architectural/code) in the project.
Then I'd end with a systems design question that takes the form of a conversation.
#1 checks you know how to code in the language you're getting the job in
#2 checks you have the ability solve harder problems and that you can communicate those ideas to other developers
#3 checks you have abstract problem solving skills and you know how to apply them to the real world
Somehow hiring people can't figure this out
Do you offer mock technical interviews by chance?
You can book him on his website for a price.
F- cj just reached 100k. We rly needa gatekeep his content. It’s too valuable
Thanks, Jesus!
Something most devs knew for a long long time...its just the tech bros following big tech. If big tech does it, it must be right.
how come a doctors are not asked to take home quizzes on how to solve brain surgery questions before joining a hospital and they walk out with 400k a year
Because they spend about 10 odd years in school before they can practice any form of medicine let alone brain surgery.
hackerrank with a big brain meme moment
That's byteboard's whole business lol
this is brilliant
Hackerrank test cases are hard to pass.. Eventhough the logic is simpler than leetcode
I don't see COBOL or Fortran, I think that's ageism.
😂 there will be a bunch of new courses like how to implement 10 features in 1 hour coming up
AHAHAHAHA lol true
Just admit it. The industry is cooked. Think about it from a business perspective: if you had millions of contracts, would you make it easier or harder to hire people? EASIER because at the end of the day, you need one million people (for example) to complete those contracts. It becomes a numbers game. You dont give a crap if someone can answer technical questions as long as they can do the job and fulfill the contracts. The problem is, companies dont have millions of contracts. They maybe have 3 big ones and that's it.
TL;DR Companies don't have enough jobs to hand out.
spent too much time learning a few leetcode problems, hope this video title is exaggerated
Adressing real learning
LOL this is the dumbest shit.
HackerRank's take on Google's whiteboard interviews must be "once keyboards are invented they will revolutionize interviews".
This debate will never end…
Leetcode has never been about testing software engineering skills. It has always been about testing IQ and it’s pretty good at that. It’s also a good way to test how much a candidate has “grinded” for a position.
Now are those 2 good measures of how well a candidate will perform? Maybe, maybe not, but what is certain is that there is no way to test that out in a simple interview process of 3-8 meetings. Their new “fix” isn’t a magical solution either.
IQ? Are you joking? Solving leetcode has nothing to do with IQ, it is all about memorizing the answers, thats it. Same as entrance exams etc. It favors people who memorize, not the smart people.
@@2A9D8F And who can learn more problems faster / understand and memorize harder problems? People with a higher IQ...
If it takes you 20 minutes to understand an easy you're gonna cover less problems than the guy who covers it in 30 seconds
Same with entrance exams: those who succeed are either smart or have grinded hard asf. I'm fine with hiring either.
@@2A9D8F are you gonna cry? jeez man
lol tell that to ppl who got lucky that they got asked a similar leetcode they practiced the day before. Also meta just makes you memorize the top leetcode Meta questions. Lets stop pretending that devoting your spare time to grind leetcode and having no life means you have a high IQ.
@2A9D8F it's not exactly memorization since you don't get the same problems, it's more like pattern recognition which is a very useful skill.
This is unnecessary Ai can't even solve a CP question efficiently and past All the edge case's i love leetcode style interviews
Havn't watched the video, but I grew to really enjoy leetcode and like it more as an interview process than an interview where I have to implement some feature or fix bugs on the spot. To me, that is much harder because its harder to directly prepare for that process, while with leetcode there is a direct roadmap and process to prepare for interviews and with enough time/hard work it is possible to master interviews.
so hacker rank realized they have lost the market and now they are saying "oh this is so bad". L take.
gg on 100k subs
now that's cool
interesting
who the hell even came up with Leetcodes
People be hating leetcode then join a workplace and write shitty code. Just because your code works, doesn’t mean it’s correct. And if you can’t thing when your problem has a very very limited scope, there is no way you are going to be comfortable working in a huge codebase. Stop fooling yourself, there is no way around leetcode. All these videos are a clickbait to attract losers who think they are too good to practice leetcode
LeetCode can definitely help with problem-solving skills, but it's not the only thing that makes a good developer. Real-world coding is about writing maintainable, efficient code that solves actual business problems. Just because someone doesn't excel at LeetCode doesn't mean they can't write good code. Also, working in a large codebase requires more than just algorithmic thinking - it’s about collaboration, design patterns, and being pragmatic with your solutions. LeetCode is one tool, but it's not the end-all be-all.
Leetcode has a very limited scope but any somewhat capable dev can brush up on data structures and algs and grind out neetcode 150 in a few months. Doesn't mean they can get a system running that actually adds value to a business.
@@4sxS307cAW True af.
I have a decently paid job in a bank and I've literally never done a leetcode problem except in the actual interview to get my job.
On that actual job 99% of my effort is spent on communication, systems design, architecture, planning, product, database, operating systems, concurrency, testing, logging, deployment or literally any else that is not related to how to find whatever in a bipartite graph.
Why do you appear angry in all your videos?
Weird intro bro
Bruh what's with the hate for leetcode questions. I get that memorizing questions is useless for an SE job.
But if that's what you feel is the problem. Just make the problems harder.
Solving hard competitive programming style problems is a really good test of someone's problem solving skills.
And if someone has good problems solving skills, (and really good implementation skills thanks to coding up algorithms you come up with to solve hard problems). I can guarantee you that person will learn way faster than a candidate who only does normal development, and fix bugs that they encounter.
Agreed. People act as if being good at Leetcode is useless. But it teaches you how to approach problems which can be applied to anything.
Most companies problems aren't really about how to structure data and most companies have little to no interest in improving performance at a faster rate than what they currently have.
But almost all companies have UX issues in their design, and features that while seemingly simple, actually affect more of the product than we realize. Some companies even have extremely basic UX, design issues and missing features and it's worth spending the time in the interview verifying if they will be able to contribute to it for sure rather than testing how they can solve a problem that is essentially a single function that runs in the context of a single file.
Leetcode is better
Back to India it is pajeet
@@baseddepartment1306 clearly you're not that based and resorted to become rasist. Time to stick with whatever job you have buddy
this sucks.
Why the hate in leetcode? It's great fun and challenging. Leetcode shouldn't be used to memeorize a problem but understand how problems work and common approaches to them and it becomes a natural way to break down complex problems and is universal knowledge not just web dev.
100k subscribers man, congrats.
Here since little willy🫶🏻
shiee cuh good video
So great that hundreds of hours poured into leetcode will be for nothing...