Skip to main content

    Whiteboard Coding Interviews: A 6 Step Process to Solve Any Problem

    By David Yang

    Reacto thumb no logo

    In this video, CEO and co-founder David Yang, spells out a tried and true method for solving whiteboard coding problems that are often used in technical interviews. Check out the full transcript below.

    So one thing I want to talk about today is how to ace that technical interview. And this is really one of the hazing processes of getting a job in tech is you got to get into a room with another developer, someone you've just met, they're going introduce you, introduce themselves to you and they're going to give you a problem and you're going have to solve it on a whiteboard which is totally unnatural, I get it.

    Usually you don't code on a whiteboard. Usually you're working with people you know. So it's a pretty intimidating process for entry-level developers, actually to be honest it's intimidating for everyone.

    It's one of those things where even to this day when I'm interviewing for technical jobs I still get nervous about it, even though I've done hundreds of them, I've been on both sides of the table and I've also, I've practiced a lot of coding problems throughout my lifetime, it still makes me nervous. So that's very natural. I wouldn't be upset that you're nervous about it, it's just something to practice for.

    So at Fullstack we've invented a method that we like to call the REACTO approach to solving interview questions. So it's R-E-A-C-T-O, nothing to do with the REACT framework it's just something that as I was coming up with an acronym I wrote out kind of the exact steps and REACTO was the acronym that came out.

    786ea6bd

    So what does REACTO stand for? Repeat, give examples, describe your approach, implement the code, test the code, and then talk about optimization.

    R: Repeat the Question

    All right so let's start with R, repeat the question. You wouldn't believe how many times I'm in the interview room with a candidate and I'm excited for the candidate and we're having a good chat and then we get to the programming question. I say implement this thing and the first thing they do is they take the cap off the pen and they start writing on the whiteboard, before they even really understood the question.

    I can tell because the first step they say is makes me just know they're going to go down the wrong path. Now one secret is that when you're interviewing with somebody people like to use the questions they're comfortable with so very likely they've done this question 100 times and this is the first time they're seeing it. So they already know all the paths that you can go down that are going to be incorrect. So you're going down the wrong path immediately and that immediately puts the whole interview onto a tough foot.

    So the first thing you should do is really just repeat the question and that will oftentimes help A, help you understand it better and B, really make sure that the things that you don't understand you can work out with the interviewer as you're going through the repeating of the question.

    E: Write Out Examples

    The second thing is to start writing out various examples. So before you even write out any code write out, "hey here are the parameters that I'm going to pass into my function and what did I expect to get out," right?

    So just create a little table, a two-column table and say given these parameters what output do you expect. This is another way to really understand what exactly the question is asking because you oftentimes will get a lot of understanding by just understanding input, output, input, output.

    One way to think about this is that this problem is a black box. How can you start understanding what's inside the black box before you know giving something into the black box, what comes out on the other side? So list out examples. This is another great way to keep the conversation flowing and make sure that you understand what the question is asking.

    A: Describe Your Approaches

    Now the third step A, describe your approaches. So start talking about the approaches you're going to take. Are you going to look for a recursive solution, an iterative solution, a heuristic solution? Give the interviewer some insight into your logic and thought process and really help them understand kind of the track that you're going down so that you know you also get a little bit of read on if you're on the right path by how they respond to your questions.

    So if you say you know what? I'm going take a recursive approach and they look at you like excuse me? Then you know that you're probably on the wrong path right? So if you say well I'm going to take a recursive approach, I'm like okay, that makes sense, then you're playing the interviewer a little bit as well as playing the question. So take some time to talk about your approach. Again, oftentimes people give the advice to keep talking, keep talking at your interviews and what they really mean is this kind of stuff, right? Keep talking about your thinking not what's going on in the code and you realize there's three things that you've done. You've probably spent about five minutes and you haven't taken the marker out of the case and started writing on the board.

    All right, so after you've talked about your approaches and you've confirmed an approach, you say this is what I'm going to take, usually the interviewer will give you some sense that you're on the right path. Then you start writing on the code going down the path of writing code.

    C: Write Your Code

    So C in REACTO is for code and you know there's so many things I can say about code, we can dive into those further if you have questions leave a comment below. I'd love to talk about how do you actually code these different problems. You have all kinds of different thoughts and ways to break down a problem. But one thing that I really like and I heard this tip from the author of Cracking the Coding Interview, Gayle McDowell herself, and she calls this Breadth-first Coding. So I'll give you a link to the talk she gave at Fullstack on Breadth-First Coding.

    And you'll hear it's a little bit of a hack right? Because you know in programming we can, we like to decompose our work into functions right? And so exactly the same thing here, decompose your approach down into the high-level functions instead of into the reality of what I say. More like the imperative things that you're doing.

    So for example, instead of writing out a for loop to iterate over some part of an array or to iterate over some data structure just write iterate data structure parentheses the data structure comma whatever, right? And then the idea here is that you're breaking down from a high level your approach into smaller steps right? So we call this idea breath first coding and I'll tell you a little bit why I think this is really an amazing hack is that really a lot of times, you can get really caught up on something that's trivial, but is kind of quite complex in the minutia.

    And so those can be things like a regex right? You can really spend a lot of time debugging a regex. A lot of times when I'm working on a regular expression, I'll spend a lot of time before I get it right and you can really get caught up on that stuff. You can really get caught up on loop boundaries. You can really get caught up on the different naming of variables and these are things that it's just not worth spending the time on when you're still thinking about the problem from a high level down to a low level.

    Oftentimes if your approach is right your core algorithm is right and the fact that you wrote out, "hey, this step is check the regex of this string," the interviewer won't even ask you to go implement it because they kind of assume that as long as they understood that this is what they need to do at this point, I'm okay with that. The underlying details aren't that important.

    Leave Room Between Each Line

    All right so after you've done the coding some additional tips, right? Leave yourself room between each line for edits right? Writing on a whiteboard is not coding in VSCode or coding in Sublime Text, it's really hard to edit. You can't really move stuff around so write out big and with a lot of space in between each one because most likely you're gonna go back and erase, you're going to insert some lines and once you start writing like insert here arrow and a small little thing, small code everywhere it gets hard for you to understand and hard for the interviewer to understand. It really just starts becoming a mess, a jumble to work on.

    Show Your Confidence Level

    The second thing I really like is star things that you're not sure about right? So as you're going through there will be some things where you're very confident and some things where it's tickling the back of your mind that you know what I should go back and check this. I'm going to just say it here and you know leave a star and say you know what, I'm pretty sure this is how this function works or I'm pretty sure these are the parameters of array reduce or I'm pretty sure, whatever you're pretty sure about leave a star there, you can go back. This shows that your confidence level matches to the interviewer's understanding of your confidence level in what you're saying. Oftentimes if you're semi sure about something and you don't convey that you're only semi sure about it, it might come off as okay this person doesn't know what they're talking about. So always try to signal to the interviewer your confidence level in what you're saying.

    Use Pseudocode

    And really a lot of times it's okay to use pseudocode in the interview. Most good companies or most good interviewers are not there to see that you know your JavaScript syntax. They're not there to see that you've memorized most of the developer network. They want to see how you think. They want to see how you break down a problem and so it's okay to use pseudocode. And so for example if there's something, just a section of code or maybe you've broken down the code into these three major things and this one thing, this one function you can't really figure out the syntax for it, feel free to just pseudocode it out. Just write it out in English and you'll get back to them about it.

    T: Test the code

    All right so the 5th step is to test, right? So after you've written your code, you're feeling relieved. You're like okay I got something on the board. Now it's very important to show your diligence and go back and say well let's take some of those examples I wrote down earlier and run them through my code. So choose a set of parameters and just run through.

    And one thing I like to do is get another colored marker and just kinda mark your variables as you're going through the process, right? So you're saying okay, A, here is one, this loop, I, is zero as I'm going through the initial one and then I'm going through the initial loop, here's the different steps that I'm thinking it's taking and then here's what it's changing, et cetera, et cetera, right?

    So go through and test your code. Run the examples and see if they actually work. This is a really important step because if you take this step then most likely unless your algorithm is perfect right off the bat most likely the interviewer is seeing some problems that they want to see if you can also see those problems.

    O: Optimization

    All right and the final step, and this is the most important, all right this can be one of those ones that really should differentiate you from the rest of the pack, right? You've repeated the question, you've given your examples, you've described your approach, you've written the code, you've tested at least on the whiteboard, written out your variables, the last thing is the O and this is what I call the optimization and runtime.

    Optimization. This is oftentimes the thing that especially beginners are a little confused about, but this is really the idea of how will this perform for large datasets? Or how does this algorithm that you've created perform under different kinds of datasets, different sizes of data?

    So this is where we hear a lot about the big O, this is big O of N, big O of N squared, big O of M times N and maybe you went with a brute force approach and maybe you went with the iterative approach so this is more you know getting into approximating big O is a little bit more detailed than I want to get to in this talk but you really want, as you know more about big O just specify here's what I'm thinking about optimization and here's what I think, I think this loop and this inner loop is gonna give me an N squared. It just shows the maturity that you're willing to, or you're able to think at a level of if I put this code into production what's going happen to our code base?

    So you know why did we think of the REACTO framework?

    And I think the first thing, the most important thing is to not start writing right when you get the code right? Because here's my picture of it is that your brain is working on multiple levels and as soon as you start writing you start committing yourself into paths that could be incorrect.

    But the more you really buy yourself time, the more time your brain gets to load in all the stuff you've practiced on, you've trained on so give your brain some time to think, right? Give your subconscious brain some time to load stuff into memory before you start writing because writing, once you start writing I feel like you start committing down pathways.

    Two, you really want start the conversation because once the conversation is flowing between you and the interviewer it's a lot easier to continue the conversation once you're coding right? Because then you can say things like well I'm a little bit stuck here, but if I think back to the approach I described earlier I think it would look something like I need to somehow find a base case or I need to figure out my recursive step. So the conversation continues to flow once it's started.

    So some additional advice is one, just stay calm, right? You're not supposed to know the answer immediately because these problems are meant to take some time to take some thought and to work through.

    So focus on breaking down the problem and I think these problems are a little bit unfair because, for example one of my favorite examples is the find me a cycle in a link list, right? So you have a link list, see if there's a loop in it. And the algorithm is named after somebody because I think they wrote about it in their PhD or they wrote about it in some kind of thesis paper. So if you don't know the answer or if you don't know that trick it's unlikely that in a situation where you're nervous you're going to come up with it on the spot.

    So it is a little bit of a test: not that you're some genius and can figure out all of computer science in a 30-minute interview, but that you've been exposed to this information. Because even after being exposed you still have to kind of recall it, implement it, execute it, and so the reality is: don't stress out that you have to figure out these things.

    If you don't know the answer sometimes that's okay. You can still get a job, you can still get to a job even if you don't get the answer and make it a conversation. Ask questions when you get stuck.

    I think the, this is just a human nature thing. It's very awkward to be in a room with somebody and no one's talking. If you ever want to do a social experiment just go on the street, someone asks you something and just look at them and as you're looking at them you're gonna see them continually get uncomfortable. They start shifting around, they're going start looking different directions because silence is very awkward.

    So this is not for you because oftentimes you're the one, you're lost in thought right? But your interviewer is sitting there like whoa, I don't want to work with this person because every time they're stuck they're just sitting there making me very uncomfortable, right?

    Whether or not this is fair or not, your job is to also make the interviewer comfortable and just keep talking, keep trying to verbalize what's going on in your mind. Even if what's in your mind you feel a little bit is totally off the mark, right?

    You have to kind of let that conversation flow and learn to read the interviewer a little bit and how they're responding to your thoughts. And final thing just to repeat, don't start writing right away. It's a reflex, you want to you know maybe you want to buy yourself time or I don't know why people start writing right away but the number one thing to remember is don't let your hands touch the marker.

    Repeat, give examples, describe your approach before you even touch the marker. And really this is something that our students at Fullstack practice a lot and the ones who practice a lot get good at it.

    They come back and tell me that it works like clockwork. You're in an interview, you're thinking to yourself REACTO, REACTO, REACTO, REACTO, react oh, right?

    I hope this helps and as always you know the most important thing is to practice too. So there's a lot of practice interview platforms, get together with a friend, and really just go through the REACTO framework. Practice it and just as a caveat don't go in there and say oh hey interviewer, I'm using the REACTO framework.

    I'm going to repeat the question. Make it natural. No one wants to feel like they're part of a script, right? So make it feel natural but this is what you should be thinking inside your head is the things that I'm gonna go through.

    So one last time, repeat the question, list your examples, describe your approach, code, test your examples, and then talk about optimization. All right I hope that helps.