generally one interview has 45 minutes of time . bro solved questions of all MAANG companies in 45 minutes. bro is a programming legend !!!!!!! orz

Truly amazing! I hope you do more of these and It would be nice if you could also post your thought process in this question and how you found the solution!

chad Competitive programmer vs virgin leetcoders ! Nice video, it was cool to see you second guessing yourself on such relatively easy questions cuz you dont trust yourself, Happens to the best of us! Keep up with the awesome content lately!

As an actual hiring manager for a FAANG-like (AR +$40B) corporation, you would sail through. Why? Because it's more important to watch your initial approach and how you transition into solving and coding. In most interviews we would even stop you and move on once we see that you have the 'chops' to eventually arrive at something that would work. ie; you're on the right track. The big plus was your attitude when analyzing the question. No real judgment, no 'snippiness' about wording, etc.. I've had candidates actually say out loud "This question is stupid", "This isn't a fair question", "I can't work on this without music", etc.. Lastly, we like it when there is a compile (or even better, runtime) error. We get to see how you react and where your troubleshooting skills take you first. For claiming that you are without interview experience, you did fantastic, IMO.

Programming is a small part of a job as a developer. You'll spend, especially as you grow in your career, vastly more time listening to others and communicating with them about the project, their design requirements, things you uncover as the project progresses, and the challenges you end up facing. The coding interview test is mainly, can you code reasonably well in a small contrived example. These are easily gamed by practicing a few core concepts that you're likely to encounter. As an interviewer I value being able to communicate what and how you're solving something far more than can you write the perfect solution. And you'll get huge bonus points if you understand the industry and the business goals driving the project.

please keep the leetcode hards incoming as much as you can. this helps a lot

Thanks Colin! Amazing work!!!

I love this video. Btw, interviewers expect us to speak out our minds a lot during the interview. It gets really awkward if we think silently.

This motivates me to code more you are awesome brother 💯❤

Awesome video man! I hope it goes more viral for you! It was very interesting and educational.

Things you don't want to do in an interview from a style perspective: 1. calling class members "global variables"; 2. creating unnecessary public members; 3. copying heavy params around (they are passed by ref for a reason); 4. using C-style arrays in C++ (one std::array cries every time you do); 5. including and using namespace std (namespace pollution and slower compilation times); 6. vector (just don't); 7. VeryLongTypeName* = new VeryLongTypeName() (use auto instead); 8. Not using const wherever you can. Overall your coding style clearly shows that you are a competitive enthusiast with not much industry experience (at least not in C++). That is generally fine for internships and entry-level positions as long as you solve the problems, but would be considered a downside for anything beyond that (at least in some companies). Other thoughts: - In q1 I would consider using dsu (not as a primary solution, but as a potential follow-up). As a data structure, it's more adaptable, in case you want to switch it to semi- or fully-dynamic connectivity later. It also reduces code duplication on the project scale that you have for dfs (unless you use visitor pattern, which I assume you are not familiar with anyway). - In q2 it's debatable whether you should've gone for the merge sort. What if you want to keep the original lists too? What if you don't want to, but you don't own them and your coworker (who owns them) wants to? Why are we using raw pointers in 2022 anyway? I would not select a policy on "in-place vs not in-place" without explicitly talking it through with the interviewer. - In q3 for even x you get that 2n is divisible by x, so you can actually do better in O(divisors), but coding it from scratch is rather painful (unless you want to resort to sieve which kind of denies the point). - In q4 using vector for static set/map is actually a competitive meta for squeezing extra logs in where setters don't really want you to (n sqrt log or ICPC being common scenarios).

I highly recommend submitting your solution 3x. The leetcode time and complexity results have been shown to give WILDLY variable results, so the most important part is just complexity which you nailed

For the facebook one: I worked on a very similar problem to this when i was in school and I think you still got it faster than I ever would have

Good stuff. Thanks for making this.

Please bring back your topic streams dude. btw nice idea

Impressive , very satisfying to watch.

I've got mad respect for people who are good at competitive programming, but it's definitely a different skill and not the same thing as being a good professional. It's possible to be both a good professional and fast competitive programmer, but most people pick one or the other to specialize in. I think the two hardest things to master are complex algorithms (including efficiency) and software architecture. That's not accounting for the low-level computer science and knowing arcane things like how to avoid L2 cache misses and optimize instructions. I focus a lot on architecture and design patterns for large, complex software and some of the low-level computer science, but I don't know tons of algorithms off the top of my head or implement them in seconds, haha. When a deep, algorithmic approach is necessary to solve a problem or there is one that applies to something I'm doing I can go read how it works and get it done ... but since that doesn't happen really frequently I've never become a fantastic student of algorithms and problems.

Love your vids man

This guy looks and sounds exactly like what i'd expect a "competitive programmer" to be. No idea why this was in my recommendations but it's all chinese to me and i envy you

this is great... want to do things like you... good video, keep it up!

Would love to see more indepth explainations for the solutions.

I'm so happy to see that even fast coders do not get their tasks, because for me looking at this was like "wtf, they just start ahead writing code when I'm still reading, they are probably geniuses", and now I can see that we're all just people

Dont know why this popped up in my feed, but it's the exact kind of questions I hate to see in interviews. I hated them as a candidate and I hate them as an interviewer these days. It tells me nothing about an engineer's ability other than they are good at puzzles. I could just as easily hand out a bunch of sudokus instead. If you've never done sudokus you'll struggle, and it might be impossible if I give you a harder one. But if you have sudoku experience, you'll breeze past them. Either way, I gain nothing of value regarding actual software engineering. The typical corporate justification here is something along the lines of "But it helps interviewers see how you approach unfamiliar problems". Bullshit. I'm not hiring you for how good you are at tackling the 1% of esoteric issues you'll encounter, I want to know how well you work 99% of the time, with the normal boring stuff. Code style, structure, readability and problem modelling are far more important that working out a solution. If I don't understand something, I ask for help, and as a team, people can always come up with the solution. But it's up to the individual engineer to actually write functional, understandable, robust code.

Where is the simulation where stakeholders ask how to connect excel to Kafka?

isn't monotonic stack better for the last question, since we can get the overall complexity down to O(n)

Great insight!

For the second problem, all numbers are between -10^4 and +10^4. Since the range is twice K, you can just map all the numbers to the range [0, 2x10^4] and do a Counting Sort in linear time and map them back before returning the result.

Hey man! You are so cool, and the video is dope. BTW I like your hairstyle LOL

Please make topic streams for med-hard graph and dp problems

Do have a video explaining all your solutions in this video? Edit: Never mind, I saw the follow up to this video in your profile video list, and the video is entitled “Dissecting FAANG Interview Questions”. Thanks, Colin Galen! I only understood one of these problems, so I’m a newb, yet I’m somehow still a professional programmer. I guess I’m just not very good with algorithms. That’s why I’m watching videos like this :). Have you considered putting a course together and selling it? You could use it as a side hustle and probably take in some decent cash flow! I’d definitely be willing to be $300 bucks for a beginner to hard level Leet code video. I don’t the language you do it matters, but I’d stick to either JavaScript, JAVA, or Python since they are extremely popular. Alternatively, you could just offer an alternative solutions for the different languages and do it in your language of choice. It would be a ton of work either way, but potentially worth it!

solving hard question under 15 mins. you would easily pass the coding interview in MAANG. Maybe the only one would give you problem is amazon cause they have Leadership Principle interview and it is a mixed bag

Thankyou for helping front and header

#3 I have no maths background so I would just iterate through all possible solutions, but hey at least I would stop when x (x being the start) is bigger than n/2 so it is at least optimised a little bit. Am I hired?

Should make more of these lol. The views seem to be popping too.

What were ur roadmap to become so good ? Can u do a video of this ?

Great content! I think a fun next step would be having a calibrated FAANG interviewer to do a mock interview with you. The biggest difference between cp and interviews is not about optimality. To be honest, you can actually clear interviews without finding the most optimal solution for every problem. A bunch of things matter on a real interview, to mention a few: - You need to be able to fully explain your approach before writing any code. Some interviewers will stop you when you start writing code to think, but some will just write it down as a red flag. - You need to create proper variable names and think more about code readability. - You don't absolutely need, but it helps a lot to explain your code out loud as you write it (if you don't, it looks like you just memorized things like the dx and dy vector to check all directions). - There is no auto-testing, and you would have to run your code manually out loud (which takes time), and it would be much harder to find the bugs on Q1 Don't get me wrong, I'm not implying you can't do these things. It is just that if you don't, then an interview question is just an easy CP question, because often the interview questions will avoid any kind of novelty to avoid bombing people who couldn't get a "trick" or something like that. Looking forward to watch more videos from you evolving on this!

is writing low-level code that that mutates anything and everything - and notusing the STL - mandatory for these things?

Your hair is marvellous, oh my god… I feel like Patrick Bateman staring at Paul Allen card

Those competitive programming sites kind of act like test prep for interviews. So that part is helpful..

Can someone explain to my simple mind the rationale of the Amazon solution, namely the logic according to which ans is incremented and why is it correct?

@Colin Galen, I want to be as good as you are, may I ask what you do in order to continue being top .1%? I have ~10 hrs a day to study what do you suggets I focus on in order to be come the best?

You didn't call delete on the ListNode* that you created on the second test. You did amazing though.

This is really helpful

I think you also want to talk about the thought process.

That's great but it's not really how interviews work. As someone that's been working for 13 years I've been on both sides a hundred times. In my opinion the two things that matter the most are 1. the approach to the problem and the thought process (perspective) and 2. being a good fit with the team (soft skills). If either one is missing you are out. Hope this helps, cheers.

If NileRed were a programmer

Hello idk if a legend like yourself would answer to a noob like me but would any Backtracking Algorithm be a good choice in the first task with the Island and do you often use any known algorithm or data structure or you have zero knowledge but invent everything by yourself in seconds or you have the knowledge but it isn't always the best choice so you invent everything in seconds?

The merge sort algorithm doesn't have the same time complexity as your solution. You are essentially using heap sort on the entire input which is \theta (n log n). The "merge sort" algorithm would actually make use of the fact that the lists are already sorted and maintain a priority queue of size at most k, so its time complexity would be \theta (n log k). Edited to add: I was talking about the Apple interview problem.

Its a really great idea to start making some interview related content. The amount of people that are interested in cp are far less as compared to interview. I wish you success and fortune. Following you from a year, best cp utuber (after william lin ( he is insane ) xD).

"UGH, why is this Java?" *switches to c++* what a hero

With the Amazon problem, the solution is also just equal to the number of odd factors of the number. So I was thinking just divide out by 2 as many times as you can, return 1 if you just have 1, or otherwise check for odd factors up to the square root, double it, and add 1 if the square root is also a factor (ie the number is square). I'm not a programmer but I do study Maths and that is what I came up with.

Wait, that algo expert guy will call you for collaboration.

Awesome video, Colin! (YOOO FIRST COMMENT?!?!)

Hey, does anyone know what is the keyboard in this video?

You're the real coding beauty!

Can you ask please your followers if they know a book or multiple books that are considered gold standard for learning all the logic necessary in type theory that is used in different papers on gradual types, dependent types? Thank you

I had plenty of interview that asked me to grab a pencil and paper and write a code there and then complained that it would not compile because i forgot semicolon.

Do workthroughs of the Jane Street Monthly Puzzles!

Codewizard but uses a microwave as a webcam.

So... This is the guy my gf told me not to worry about... You're damn good bro

24:45: I don't understand how a mergesort solution is going to be better. The naive mergesort solution would just be to dump all the linked lists into an array and mergesort that, but that's O(n*log n) and still takes O(n), strictly worse than using a priority queue merge. The slightly better option is to append all the linked lists together and merge those together, which can be done in place, but is still O(n*log n), which is worse than the O(n*log k) solution. The only other idea I can think of that uses mergesort is to mergesort the heads of the k lists to find the smallest one, and repeat that n times, but that is O(nk*log k) and O(k) memory, so still strictly worse than the O(n*log k) solution.

All of that in 3 years? ZAMN!

q1 O(n^2 - area of all islands) is pos, you only need to find the perimeters of the islands to calculate the area, and then just find the best grid of a perimeter whose islands shared have the max sum area q2 u could just O(n > 10^4 ? n : 10^4) da shi by doing counting sort

Absolute Chad programmer!!

Thank you

The constant ctrl-s is so relatable...

Cool video! One aspect of interviews that is greatly different from competitive programming is the fact you cannot test run your code. You need to "whiteboard" test your implementation and use clever test cases to spot the problems quickly. Of course knowing how to solve the question is like 80% of the work but communication, testing and overall organization is really important too and those can get you into no hire zone if ignored.

Please Can you make a video on how to get interest in coding

very informative

Nice hair type brotherman

Is Q1 a trick question? The specifications don't actually require the program to make a large island, merely to [optionally] make a single node change, and then return the size of the biggest island. The simplest solution would be to make no (or a random) change, and run a graph segmentation. Or is this question just REALLY badly written? E.g : "Example 1, Output: 1, Explanation: change one 1 to a zero, leaving exactly one island of size 1." This fulfills the requirements as written. If this is not what they actually want, who-ever wrote the question fails at writing good product specifications. Should the requirements be: "return the size of the largest island _possible_ in grid after applying this operation."?

Are those who ace the faang interviews considered vampire slayers? It’s morbin time?

your hair is so so pretty

Good for coding interviews

Two things attracted me to this video : 1st Relevance of topic , 2nd I thought there was a beautiful girl in video from thumbnail. I was like "Ek dum se jasbaat badal diye "

Make video on how to learn minimum maths to be good at dsa?

This is the new standard to get into a dev position, unless of course you know people that can get you in. In that case you don't even need to know how to code.

Is it possible to learn this power?

I like your thinking, but as you yourself know and said, not having much industry experience shows :) a few little comments of the things you're talking while writing the code are nice, though many of those might be redundant once you look at the code,... 5 years later someone might stumble over the code because of a problem and not having the original problem :) so some guidance into the "thinking" of the code is always helpful in any case for an entry level position in any language you're so massively overqualified, I wouldn't say no if you applied to our company and then sit in my classroom for me to teach you ABAP xD

I believe it is possible to solve the final problem in O(N), if you keep a stack of strictly increasing (from top to bottom) elements for odd-numbered jumps, and the opposite for even-numbered jumps.

Woah, I didn't know BrutalMoose had another channel

Naur an edraith ammen! Naur dan i ngaurhoth!

What are these loops? It looks like a maschine was coding that loop! Crazy this dude is crazy. Thing I would want to know can he program all the solutions in other languages? Like Java, Kotlin, Python, Typescript`?

thank god i own my own software company. going back to corp america would be detrimental to my mental health.

The sad thing is that programmers have to go through the meat grinder of corporate interviews. I hope you can get back to pure programming as soon as possible!

Apple: concatenate all the lists in one and sort?

But tbh it was easy for you) I guess competitive programmers best suit for coding interviews. It's harder for guys who can actually build stuff that works but during interview they are forced to solve isolated tasks about numbers using bare arrays that no one really uses in every day work)

Man, please make a detailed video on Dynamic programming questions from code forces rated between 1000 - 1500

For the 1st one (Making a Large Island) you would be better off with line scanning (since it's connected only in 4 directions) instead of recursive growth of the components from each point.

can u make a sheet where a beginner should start coding for competitive programming

You look like one of the Elves in LOR. That is all, I'll show myself out.

So on the real interview there would be a guy asking you a stupid questions and forcing you to explain what are you doing after each line of code

How to get so good at programming?

You'd smash the interview process (coming from a new post-grad who's just secured a job). I'm not coming from a FAANG perspective, I work for a big-data company. Securing the bigger roles is 90% coding skill and 10% people skills. And you've got both.

The first problem can be implemented quite easily and super efficiently O(n^2 * log*(n)) using disjoint-set data structure. DFS will be slower, O(n^4) worst case.

Can someone summarize the video its took long Thank you : )

I need to step my math game up...

How old are you, and at what age did you start coding?

How long before AI will answer these questions in a few seconds if not milliseconds?

Is Netflix really a big, sought after tech company? Or is it just tossed in there to make the acronym more palatable?