Take-home interviews

Seems good, I like the choice.<p>In discussions on this topic I see a lot of: &quot;Programming on the spot is hard, let people program at home&quot;! But then other people say &quot;Why should I program for free at home, my resume clearly shows I am already a skilled programmer. All this will do is cater to young people without families, or those fresh out of school&quot;.<p>I am not looking to hire devs right now, but I am thinking about the same idea with a tiny tweak when we move to that stage. If the verbal interview goes well, offer 2 choices. Make it clear neither choice is seen as the better choice, and both will be judged equally.<p>* Sit on a computer with me for an hour, and show me how to code a fairly simple program.<p>or<p>* Get a small actual work assignment to take home and code. Would expect it to take maybe 5 hours to code. Offer $500 as a 1099 contractor to complete the assignment within the next 2 weeks or so.<p>This gives both groups a chance. The too busy to do a big programming assignments can code in front of me for an hour. I should be able to judge their chops pretty fast. For the people more nervous to code in front of me, they can get paid a nominal fee to code some small piece of code.<p>The only group I exclude is the group that doesn&#x27;t feel the need to show code to land a job, but not too interested in that group. I have seen too many good talkers and bad coders to want to risk this group.

From the employees&#x27; perspective:<p>Take home tests are the worst. Company says take home test will take 3 hours to complete. They never do. Schedule 2x or 3x the estimate. Especially if you want to impress the reviewer.<p>You send it over, then the company says no or yes, only to move to new stage.<p>In the worst case you ruined your weekend and received a no. But the company just took 10 minutes to arbitrarily reject your application.<p>From the company&#x27;s perspective:<p>I&#x27;ve seen applicants receive friends&#x27;&#x2F;roommates&#x27;&#x2F;spouse&#x27;s help on take home tests. Not a good indicator at all even with a glowing submission.

This seems like a wonderful idea.<p>I hate coding interviews, because I freeze up and I look like a total idiot and cannot do even the simplest things. Same thing at exams in school or university - my mind just went blank and into a noisy self-rumination loop. I guess it&#x27;s called anxiety.<p>I interviewed with Google last week and bombed it, even though the problem was quite simple and I would have solved it wonderfully without all that anxiety.<p>My mind just can&#x27;t produce any sensible thoughts when someone&#x27;s breathing over my neck and there&#x27;s a timer. When I&#x27;m solving the problem in real time and talking about it, I have to stick with the split decisions that I verbalize and build on top of them, even though further down the road I realise they&#x27;re not optimal and this contributes to the anxiety even more.<p>It&#x27;s hard to &#x27;refactor&#x27; your ideas during a 45m coding interview, but this is exactly the process we go through when we write &#x27;real&#x27; code - we try a thing, then we improve it by refactoring, we optimize it, we find and fix corner cases, etc.<p>A take-home problem, on the other hand, would be totally cool. I have time to think about the problem, come up with solutions, optimize, unit test - do the real programming thing which I&#x27;m being hired for.<p>I would then gladly discuss and explain the code with the interviewer.<p>If then I would be given the task to extend the program with a new feature, then I would be familiar with the data structures and algorithms used and would probably find it much easier to extend, than starting with a blank file and figuring it out on the spot.<p>But I guess not everyone agrees with the home interviews - some think it&#x27;s a waste of time and I&#x27;d have to agree..<p>So I guess the optimal solution is to offer the option of on-site coding interview or a take-home problem.<p>People like me would take the problem home, build it and shine at it, others, who&#x27;s minds are sharpened by the adrenaline would take the on-site 1hr coding challenge.

Take home interviews are a great indicator of a company&#x27;s hubris. &quot;Its so awesome to work here people are going to jump at the chance to do my 3 hour homework assignment&quot;.<p>The problem with them is fundamental: &quot;You will only get someone desperate enough to take your exam.&quot;<p>If the person is qualified they will be swimming in opportunities and will likely throw your exam directly in the trash heap. If they aren&#x27;t you probably don&#x27;t want them working for you. I suppose these might work if the entire industry colluded on it but then ... prisoner&#x27;s dilemma.<p>Maybe a more useful indicator is weeding out the candidates that didn&#x27;t say &quot;no thanks.&quot; Cause really why are they so desperate for a job and why do they have all this free time? but then ... ethics.

As many other parts of an interview, I&#x27;ve always found the blackboard coding session extremely strange. When was the last time you coded in TextEdit with no docs around, no time to think, standing up, and being watched over the shoulder?

This is presented as a weeding out technique but it is actually a negotiation step. This allows the employer to establish precedent that you work off hours from home. This allows the employer to identify candidates that are willing to do whatever it takes for the job. These are the same people that won&#x27;t do hardball negotiations for salary, employment terms or working conditions. Rather than weeding out the people who will fail the test, it weeds out the people who understand that asking for homework is unreasonable.<p>My last two phone interviews ended with the recruiter &#x2F; hiring manager explaining that the next step was a take home assignement or test. I said OK, but haven&#x27;t done either one. Next time I will politely tell them that they can look at code samples in my gthub but I am not spending my time jumping through their hoops just for the chance to have an in person interview.<p>It&#x27;s bad enough that we have to burn an entire day to do an in person interview...there shouldn&#x27;t also be homework.

I like these great in concept. When I&#x27;ve been between jobs, I&#x27;ve been very happy to participate in them.<p>A couple open questions:<p>1 - Is it reasonable to expect a 10x programmer whom your are trying to poach to give up so much time? (Or should you give them a $250 Starbucks gift card or something similar for their time?)<p>2 - Can you really ferret out cheating? I had a grad school classmate who paid someone to do his (non-programming) take-home homework for a job interview, and he got the job. He only lasted 6 or 7 months, but it was enough to be awful for all parties involved. I don&#x27;t have a great counter-solution other than ask for someone to come in to the office to do the work, and even then you can&#x27;t tell if they have remote support.

One thing I always found interesting in these discussions is how the same people who are quick to cite The Mythical Man-Month when it comes to the futility of adding new people to a late project will also claim that working 1-3 hours on some trivial problem is &quot;doing free work&quot; for a company. How much value do you think your 3 hours of work with absolutely no context can possibly add to the company? A company trying to get meaningful work done by secretly farming it out to interviewees is beyond ludicrous. Interviewers are only interested in finding good hires, not in conning you into doing their homework for them.

Take home interviews are common in data science, and they are deeply exploitative. I have been asked to spend anywhere between an hour to a week on a project, and in most cases they simply decline to hire you without giving you any feedback. In the worst cases, companies like Knewton and Mattermark have given take home interviews that not only took a lot of time, but were related to the companies core business model. In other words, it&#x27;s free consulting.<p>And finally, you aren&#x27;t fooling anyone when you say that it should take 3 hours, or as long as they want to spend on it, while giving them 3 days. You are pressuring people to spend as much time on the problem as possible to &quot;prove they are a good programmer.&quot; It&#x27;s exploitative - you aren&#x27;t paying us and the probability that we will get hired is extremely small. Think about how many companies we are interviewing with.

As someone who is still technically waiting - after a month - for his phone interview, I like the idea of doing some actual coding. If you have to make people jump through hoops, at least you might try and make the hoops at least somewhat meaningful.<p>Edit: I just pressed the button on my TripleByte dashboard to switch to the &quot;project-based&quot; track, but then I get redirected straight back to the same &quot;we&#x27;re sorry&quot; page I&#x27;ve been seeing all along.

Here&#x27;s the best approach I&#x27;ve seen yet to the take home interview (and interviewing software engineering candidates in general).<p>Here&#x27;s a git repo, a problem statement and a slack channel to ask questions. You can use any tools you like and spend as much time as you like on the project for the next week.<p>----<p>Employer&#x27;s perspective: You get to see some code, see how they approach a problem, see how they use their tools, and get a sense for what it&#x27;s like to work with them.<p>Candidates&#x27;s Perspective: You get to use the tools you&#x27;re already comfortable with, you can set your own pace, and you get a sense of what it&#x27;s like to work with the team.

Take-home interviews are a <i>perfect</i> match for startups - they really reflect actual working conditions. One way or another, you&#x27;ll be taking home your work every day! :)

We (Sauce Labs, Mobile Team) are experimenting with take-home tests as a first-contact screen. I&#x27;m familiar with the take home exam that takes a week to solve, so we went in a different direction.<p>We schedule time with candidates, send it to them at the right moment, and expect them to send in a solution two hours later.<p>Before we tried it on candidates, we all did our own test, and we confirmed it could be done in less than two hours. (Granted, we are very familiar with the problems we care about, but we usually get complete solutions from candidates.)<p>So far the results have been fascinating. The test is conceptually super easy, deliberately so, but with a very wide set of possible implementations. It is more about tying together moving parts in the right design. The goal was to do the opposite of the algorithmic brainteasers you normally get in interviews -- it&#x27;s a miniaturized version of the problems we actually solve all day. Hopefully, it&#x27;s a good test of what kind of coder the candidate is.<p>But we&#x27;re still working on it and refining it.

The problem with random or unaligned assignments is you end up with a randomly skilled programmer. Often not bad, but not on point.<p>The problem with Rosalind or Project Euler assignments is you end up with an excellent theoretical math or bioinformatics programmer. Often not bad, but not on point.<p>Fundamentally anyone with a degree or experience or a non-trivial github can write code, but you want to test their judgement, their thought process, their comprehension, their style, their knowledge about your business domain. Other than total open field blank slate projects (very little of my time over the last 35 yrs has been spent in that mode) you usually have existing systems and code. So give them a &quot;special&quot; sample of your own code. Shouldn&#x27;t be too hard to find unless you&#x27;re literally hiring the first technical employee. Then give it to them a couple days before an interview and inform then you&#x27;re gonna review that code together, they&#x27;ll present you with a rewritten, redesigned version, and then review the rewrite together.<p>If you&#x27;ll feel bad about making them do &quot;real&quot; work, the best code to send out is some that has been heavily customized to only work some of the time, not properly error check, and intentionally somewhat obfuscated, so I sincerely hope that code thats screwed up to that level has to be intentionally manually generated for the interview. So strip out most of the failure&#x2F;error detection code, screw up some of the code, maybe intentionally cut and paste an almost identical function in place to see if they clean that up. Some folks like intentional outright errors, like typos, is this the kind of programmer who can&#x27;t write English? Also wipe most of the comments, put some intentional logic errors in the comments. This can be fun...<p>If you&#x27;re looking for non-intro level programmers &quot;everyone knows everyone&quot; and my latest job we didn&#x27;t talk programming because I was vouched for as knowing what I&#x27;m doing from years of coworker experience at a past employer. I&#x27;d be moderately offended if someone I worked with for five years asked me to fizzbuzz, either in person or as a take home test.

I like this idea. I took the Triplebyte quiz mostly out of curiosity and then decided to stop when I was prompted for a phone interview. I like the idea of &quot;pre-qualifying&quot; myself for a job even before I&#x27;m ready, so that when the time is right, I can pull the trigger and switch easily.<p>With a phone interview you have to commit, which is no problem if you&#x27;re in the market. But if you&#x27;re not in the market and you still want your options to be open, interviewing on your own time makes a lot of sense and reduces the mental burden and stress of switching.

I&#x27;m not particularly a fan of take-home tests. But comparing that to mumbling at a white-board, while a software engineer is pretending to be an expert psychologist at analysing your ~train of thought~, and unfortunately, more often than not, someone in the board is only paying attention at how much of an WASP male you are or pass by it, where you need to do a task that you only repeat at interviews for a job that will not require that...compared to that, give me take home interviews any day of my life. Yes they are free work, but at least they are free work related to actual work. Live coding in the whiteboard performances are not part of the work. And the stress argument (that programmers must be able to deliver under stress) there are categories of stress, social anxiety of a live performance is not the stress we deal at our works. Sometimes I do good whiteboard coding interviews, sometimes I do bad whiteboard coding interviews, none of the situations I felt the merit was of the code or my knowledge of the subject, it is not a proof of how well a programmer does his job, is a somewhat related way of knowing if somebody knows who to code something and if he does not feel too nervous to code on a whiteboard instead of a computer to a bunch of strangers testing him. If you add that to the fact that some interviewers like be to randomly arrogant, you&#x27;re missing some really nice but perhaps shy coders.

One thing that a local company did was to give a simple programming assignment (given postgres database connection info, create simple REST API using any language). They would switch up the details of the assignment for each candidate (e.g. different endpoints for same set of tables, different tables, etc.) to combat cheating.<p>The applicant was expected to be able to answer questions about the code anyways, so hopefully that would&#x27;ve been an effective layer of protection as well. Don&#x27;t know if this company is still doing it.

Historically, I&#x27;ve tended to do poorly in in-person interviews. I felt my critical thinking and problem solving skills plummet to a fraction of what I am capable of. I initially thought that with enough real-world interviewing experience, I could familiarize myself to the stress, but that never really happened. Interviews tend to be few and far between, so the familiarity never really sticks.<p>I usually perform better on take home interviews, but 90% the time I&#x27;m unwilling to dedicate what are usually days for just a chance to be accepted at a company I may not even want to join. I think many employers use the take home interview as a screener without realizing they need to first cultivate in the candidate the enthusiasm and willingness to complete it.<p>Some ways companies can encourage me to actually complete the take home interview:<p>- Provide compelling information about the job (estimated pay&#x2F;equity, meeting the team, evaluating culture-fit, getting me excited about the problem, etc). Lots of companies save this step for the courting process that happens after the technical screen.<p>- Pay me a modest amount just to complete it, regardless if I pass or not. I don&#x27;t care about the money, but at least it won&#x27;t <i>feel</i> like wasted time.<p>- Make it way shorter, but then it might be useless.<p>--<p>What I&#x27;d prefer instead of all this interviewing is still:<p>1) meet the team, evaluate culture fit, etc.<p>2) discuss past projects in detail, maybe do a code evaluation if relevant<p>3) contract for the company for 2-4 weeks at market rate pay. Hell, make a 10-week &quot;internship&quot; out of it, whatever.<p>4) receive offer (or not)<p>This is not always practical or scalable, but I&#x27;ve gotten offers 100% of the time this way, including at YC companies and other startups. And from the company&#x27;s perspective, I think it extricates any unreasonable expectations they have from prospective employees. It&#x27;s always tempting to look for unicorns when you have hundreds of resumes to choose from, but when you have a likable contractor doing good work, there&#x27;s no reason not to give him&#x2F;her an offer. Plus, the employer can evaluate the single most relevant skill in an employee -- the ability to learn.

I really hope this works. I never understood whiteboard interviews. They don&#x27;t reflect how real development works or how that developer works either. I hope you post an update when you have enough info!

Incredible to read so many cases here where people have done a take-home exercise and heard nothing back. I had assumed that the normal practice would be to give a take-home exercise as part of prep for an in person interview, so if you do the work you&#x27;re at least guaranteed the opportunity to talk about it.

This is great. I especially love that candidates are given a choice, so if they prefer the traditional technical interview they can choose that.<p>Normalizing performance between the two interview types will be challenging, but I think the benefits to be reaped far outweigh the difficulty of the challenges.

How does this work if candidate has experience in, say, backend, but haven&#x27;t worked much on frontends or mobile apps etc? A specific &quot;take home&quot; task puts those candidates at advantage who have already worked in related area, have written code that they can immediately reuse or are experienced in frameworks that allows them to fast trek. This would be great if the job also required exact same capabilities for the foreseeable future from the candidate. However it would overrate or underrate the candidate if the job requires candidates to work on diverse set of problems in long run (for example, switching from MySQL to graph problems). It would be interesting if article also gave few examples of such &quot;take home&quot; tests.

That looks great. Are there enough problem seeds to help combat sharing of answers?<p>Now we just need you at non-YC startups, too.

&quot;Hey! Stop measuring me and give me job! I&#x27;m smart you asshole.&quot;<p>~ every software engineer in the world<p>Anyone else realize we&#x27;re the only ones who complain about this shit?

My personal favorite method is similar to this and we use it at my current company.<p>The big difference of course is that we pay for the candidates time. Not everyone has 5-20 hours to burn on a project in terms of free time. It&#x27;s a big commitment, especially for the folks with families.<p>Some key benefits of a project for our interview process:<p>* We get to see a candidates programming skills with a greenfield project. This gives us a chance to analyze all aspects of development (organizing code, creating tests, using best practices, ect.).<p>* Candidates get to act as if they are on the team, so they are encouraged to ask questions, get feedback, ect. throughout the project. This gives us a chance to see how well they communicate in addition to their programming skills, which is REALLY important to us, since we&#x27;re an entirely distributed team.<p>* The project is not throwaway for the candidate nor us. We usually have an ongoing list of libraries and tools that we need internally for our projects, so the candidate is actually contributing to our mission. <i></i> This also justifies paying a candidate for their time <i></i>

&gt; The project-based track will require a larger time commitment<p>This is the only part I take issue with. I tried a take home interview once that took the better part of a weekend and decided I&#x27;d never do it again. There just isn&#x27;t enough time. Basically, if it takes that long, there needs to be a very high chance that I&#x27;m getting hired at the end, or it&#x27;s not worth it.

I would never take this option.<p>At my current workplace, I have the tools I need to work. Most of my softwares licences are under my employer&#x27;s name. I have access to my snippets, my previous projects and many build processes (minifying my code, preprocessing my CSS, etc.). My workflow is great.<p>If you give me a project to do at home, I can&#x27;t use any of that.<p>I would have to spent hours working either from my bed, my dining table or in a coffee, on a laptop. I would also be required to work on this project after the 40h&#x2F;week that I already do.<p>I&#x27;ll take the regular interview any time.<p>One of my worst interview process was one time when they required me to do a project of 20+h in a week. They knew I was employed elsewhere, so they let me give it Monday instead of Friday. This was ridiculous.<p>Otherwise, everything was doing very well. However, the experience really discouraged me from working there.<p>But I guess it&#x27;s not all that bad, because I most likely dodged a bullet. If they can&#x27;t even give correct deadlines to future employees, I can&#x27;t imagine how they threat their own employees.

I view these as a good litmus test for the company too. If they won&#x27;t answer a question about the problem, or think it&#x27;s very self explanatory, it might mean they&#x27;re not the best at communicating.<p>I recently did a clojure problem where in the gist itself I asked for feedback and showed my thought process, but crickets until a month later when I followed up and they said they had moved on. I think it was for the best.<p>A good preventative measure might be agreeing beforehand on a time to discuss the solution. I know I certainly will next time.<p>Also hiring managers need to ACTUALLY read code. There have even been times I&#x27;ve asked if they&#x27;ve read my code before doing a lengthy drive, they&#x27;ve said yes, only to realize upon getting there they haven&#x27;t. If you won&#x27;t take some time to assess the code I&#x27;ve written and put my name on, I don&#x27;t think I should be expected to drive 3 hours to meet you.

To me the value in a take-home test is in smoke testing what the applicant said they are familiar with. We once accidentally hired someone once who talked the talk but didn&#x27;t even know SSH.<p>My test involves a basic task (that is related and already implemented in our codebase). You have to check out a skeleton project, add the solution, build it, and deploy it. The test takes me about 30 minutes and most applicants spend 60-90 on it. Everyone on the team reviews the test and the criteria is essentially &quot;do you want to work with this person&quot;. It&#x27;s very rare for someone to pass the test (which occurs between a phone screen and an in-person interview) and not get an offer. The exceptions have all been around compensation issues.

The take-home&#x2F;programming tests in general are just as bad as interview questions. You&#x27;re asking senior level engineers to take 3-5 hours (at what? 100-150$ hr bill-rate) to see if they can code?<p>Do you like them? Check. Do you feel like they are telling the truth? Check. If you use the &#x27;try before you buy&#x27; method: You can easily see if they can handle the job&#x2F;handle the pressure.<p>Also, you probably aren&#x27;t the only company&#x2F;avenue they are applying towards. All of the calls&#x2F;tests add up.<p>The question is though, since companies are already using take-home tests&#x2F;interview questions... can anyone confirm that these methods are more effective than a &#x27;try before you by&#x2F;contract to hire&#x27; position?

My old employer Thoughtworks does this.<p><a href="http:&#x2F;&#x2F;www.thoughtworks.com&#x2F;careers&#x2F;application-process" rel="nofollow">http:&#x2F;&#x2F;www.thoughtworks.com&#x2F;careers&#x2F;application-process</a><p>It&#x27;s a fantastic way to vet candidates and there are multiple exits to the process. For obviously bad submissions, it&#x27;s probably a no-hire, but I&#x27;ve even seen incomplete assignments get someone to the next level and then a hire.<p>Some people get sent to ThoughtWorks University which basically trains them to code, work in an agile environment and how to be a good consultant. Everyone that goes says it is a great experience.

In 2008, I was hired to work on a tech startup from Australia, while I was living in Brazil. After I went through an initial quiz and a few interviews, they were willing to give me a shot by giving me (and paying for) a small project to code. Only after succeeding at it, they hired me, sponsored my visa and paid for all the relocation costs. In your case, I really think this approach would be a good bet in cases that the YC company is willing to sponsor the visa from the potential employee, so giving him a paid project before even flying him, might be a good filter.

<p><pre><code> Rather than pick a new project, however, they&#x27;ll take the same project further, incorporating feedback from the 1st interview. </code></pre> Is there still judgmental coding during the 2 hour interview?

We consistently do this in Sandglaz whenever we interview for a technical position. It most closely aligns with how work is done in real life, and we give candidates a reasonable amount of time to do the project. We ask them to document what they would do if they had more time. It is more of a way of peering into how they organize code, their understanding of technical debt and prototyping, and their ability to write unit tests.<p>It&#x27;s been invaluable in determining candidate fit.

This sounds like a brilliant idea. I hope that you are also including homework assignments that cover (1) communication skills (especially WRITTEN communication) and (2) personality. In my experience, those two areas are far more critical to success than raw coding skills. Also, everyone should pass the &quot;no assholes&quot; test, if there is such a thing.<p>I hope that you expect the same amount of prerequisite work for all of your candidates (marketing, finance, sales).

Wondering if under this model an unethical person could set up a fake interview process with real assignments to get unpaid work done by others who want the job.

Tech recruiter here: I have seen so many companies that lose out on good developers because they ask them to do a take home test.<p>Obviously you want to vet the person technically but you should be able to do that through talking with them. Most developers put off take home tests even if they are excited about the company and by the time they do it they already have finals&#x2F;offers from other companies.

I actually did a similar thing as an ops guy (debugging systems as described in some questions) and really liked the concept. On the other hand I work on mathematical riddles for fun so maybe I&#x27;m the core target for those.<p>It doesn&#x27;t replace face to face interviews but I can see it as a good example of &quot;how would face this real life situation you would encounter if you got the job?&quot;

This is great, but I feel like take-home interviews are better for a first-level filter. I went through this when I applied to Thoughtworks, but it definitely wasn&#x27;t the only interview. This may have been because I was in a different state also though, so before they would pay to fly me out, they let me pick out of a few problems and then email them in within 72-hours.

Why don&#x27;t interviewers just have a conversation or is that so totally lost these days? I find it a whole lot easier to just ask the right technical questions in the first 2 mins. to decide on hiring.<p>I keep seeing more and more dis-associative non-communication being presented where the interviewer attempts to do less and less. Its lazy and a crap way to treat people.

I hate take-home interviews and rarely finish them. Especially when a candidate has many options or places a high value on actually getting to meet possible future team members (as I do), a two or three-hour assignment take-home assignment is a lot to ask. I would rather get to show previous projects and explain context and what I might do differently now.

&gt; We expect them to spend about 3 hours on the project (or as long as they want to spend to show us that they&#x27;re a good programmer).<p>3 hours isn&#x27;t too bad. I had a friend interview with some hedge fund and he said he probably spent 30+ hours on a take-home project (he didn&#x27;t get the job). I thought that was way too excessive for just 1 interview.

I did one of these when I applied at my most recent position. It took most of the weekend and actually it was a lot of fun. But I doubt I&#x27;d do it again if it required more than a few hours of my time. Any test that time consuming is arguably biased against those with families.<p>Some advice to those doing take home tests: write (good) comments and include tests.

Gauging ability to operate under pressure is valuable too...ever tried to fix a bug on a device that was supposed to ship yesterday, under the pressure of the entirety of the hierarchy above you all the way up to the CxO?<p>It helps to know if the candidate you hire will handle that, or fall apart just when you need him&#x2F;her to do the job he&#x2F;she is paid for.

This sounds like a great alternative and certainly would work well for me. What I&#x27;m afraid of though, is that I&#x27;d pass your interview and the aforementioned YCombinator companies would again want to put me through the on-site, high-pressure whiteboard technical interview.

I&#x27;ve done this, and liked the process.<p>Interview with 2 people for an hour or so. Then you get a Rails project, fix 3 bugs, choose and implement 2 features from a list, then 2 features they choose. Come back in a few days and discuss why&#x2F;how you did what you did.

The problem I find with these is that while solving a problem might take 3 hours, you can spend an essentially unlimited amount of time polishing code - writing comments&#x2F;javadocs, neatening up the code, writing more unit tests etc. etc.

Is there data on the efficacy of non traditional coding interviews vs traditional ones?<p>I&#x27;m all for improving the process, but as someone who at times has struggled on a coding interview, I believe that the onus to improve should be on me the candidate.

Why is it ok with everyone that triplebyte is using desperate job seekers as lab rats?

I considered a job out in SF working for a decently well-known guy, Andrew Chen. I had gone through 4 coding interviews progressively working my way up their ladder to their CTO. Now, each of these I did well on but they didn&#x27;t like how fast I went. Despite them saying &quot;Our fastest done was 1hr!&quot;, they didn&#x27;t seem to appreciate me going fast and came to the conclusion that I probably didn&#x27;t know how to write production code -- just a quick hacker. Actually, it was rather annoying because their &quot;top&quot; iOS dev didn&#x27;t even understand basic shit I was doing. He&#x27;d have to keep stopping me and being like wait -- why&#x27;d you do that? Yet as I worked my way up, I really enjoyed a couple of their engineers and CTO. I didn&#x27;t have to hold THEIR hand during a code interview.<p>So I was an iOS developer getting considered for Android position, and they said &quot;hey it&#x27;s great you can code fast &amp; all, but can you learn fast?&quot; and told me to create an Android app that searched images on Google. I sighed because I had already been through 4 coding interviews, but oh well, this will be it I thought. Since I wanted to do a good job, I spent two hard days working on it basically all day every day. Since I was new to Android world, I had to learn that while being productive. You can see the result of my efforts here: <a href="https:&#x2F;&#x2F;github.com&#x2F;joslinm&#x2F;android-image-search-example" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;joslinm&#x2F;android-image-search-example</a>. I thought I did a nice job because rather than using a HTTP library, I read the streams myself to give nice progress indicators for each image (which persisted even thru phone rotates).<p>I give it to them and they say, &quot;ok seems to be working.&quot; After that, I was invited to SF to interview with them in person -- actually, interview is the wrong word, I would be working with them for 2 full days. So we did that. When I wasn&#x27;t coding, I was being grilled about my work history (perfectly acceptable, but why didn&#x27;t we do this earlier?). Finally, after that, I was told they had come to their decision. Thinking I was a shoe-in, I went to go meet with Andrew. Andrew&#x27;s decision: &quot;You&#x27;re too entrepreneurial.&quot;<p>Now let&#x27;s examine this for a quick second. They had come to a pretty fair conclusion that I was too entrepreneurial after talking to me. They realized, hey this kid is a go-getter and probably isn&#x27;t the greatest fit for our #7. Yet this was AFTER copious amounts of coding and wasting my time doing all these code interviews and even joining them to code in-house. If they had just went about a pretty normal interview, they would have discovered this fact way earlier and not wasted either of our time. So why waste my time forcing me to code before they do their diligence? Because he&#x27;s Andrew Chen; you can google his name and discover who he is. It would be a &quot;honor&quot; to work for such a prestigious start-up pundit. In other words, he can get away with it.<p>For those of you considering a start-up and actually have skill.. don&#x27;t get taken advantage of by these bogus work projects. The employer wins because he&#x27;s not paying you and gets a lot of evaluation for free. And if they do insist on you working with them for a couple days, you insist on getting paid for those couple days.

I thought Gayle Laakman McDowell, who wrote &quot;Cracking the Coding Interview&quot;, wrote a good article about the problems of take-home interviews.<p><a href="http:&#x2F;&#x2F;www.gayle.com&#x2F;blog&#x2F;2013&#x2F;09&#x2F;18&#x2F;companies-who-give-candidates-homework-assignments-knock-it-off" rel="nofollow">http:&#x2F;&#x2F;www.gayle.com&#x2F;blog&#x2F;2013&#x2F;09&#x2F;18&#x2F;companies-who-give-cand...</a><p>At the core of the problem is that this approach can be used to burn a lot of a candidate&#x27;s time without an equivalent investment from the company.<p>I&#x27;ve mentioned this before - I applied (maybe 5 years ago) to a company that first asked me to take a Java test (about 1 hour of work). Then the recruiter called me and sent me a take home project that should take &quot;5-7 hours&quot;. I did this, crickets chirped for a month, though I did check in with the recruiter for a while (I gave up eventually). Finally the recruiter called me with a one-line brush off &quot;we&#x27;ve decided not to move forward at this time…&quot;<p>Granted, this was a bad experience, and it won&#x27;t always be like this. But I&#x27;m pretty close to saying &quot;never&quot; to these exams.<p>I thought Gayle&#x27;s suggestions were good. She goes so far as to suggest a 90% passage rate - that you should not be giving these tests to people who are unlikely to pass then, using them to confirm, not screen.

But do you want someone who cannot handle an interview?<p>Perhaps take-home testing works if you want a &quot;code monkey&quot; who can sit in a box churning out code according to specifications, but I cannot think of a company that doesn&#x27;t want more. The ability to interact with others in sometimes stressful environments is, imho, an essential skill for all.<p>There is also the danger that companies will turn these &quot;tests&quot; into work product. I did an interview last year (legal) that requested &quot;writing samples&quot; on very particular and timely subjects. Once I realized the game I asked them to sign a copyright agreement. They got screaming mad and tried to get me to sign all sorts of NDA junk.

I usually refuse them unless I have spoken with the technical people first. You get no idea about the job from HR staff.

&gt; Propranolol is used by musicians, actors, and public speakers for its ability to treat anxiety symptoms activated by the sympathetic nervous system.<p>- <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Propranolol#Society_and_culture" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Propranolol#Society_and_cultur...</a><p>Can ask your doctor about this

Do this instead of white boarding or other coding during the interview. Please.

I dont really understand the way interviewing is set up in general. Interviewing in my mind is a two-way street. It’s not just the candidate that’s being interviewed for the job, it’s also the employer being interviewed to determine suitability.<p>In light of this, I think most interview processes don’t make any sense. I think that a screening phone&#x2F;Skype call makes sense before either side commits any time&#x2F;resources etc. In this process, I think there should always be room for the candidate to ask questions of the employer. Any situation that is unbalanced, in which the employer is playing the role of the one with the final decision is a completely skewed setup.<p>I interviewed a while back with a YC company and although I didn’t get the job, I was highly appreciative of the setup. Firstly, there was a Skype&#x2F;phone call, which felt really like an open conversation; one in which I was free to ask questions and understand suitability too.<p>Then I was invited to an in-house hack. Essentially, I spent the day (a Saturday) hacking with the founders on a project. The project was a lot of fun and something that I learned a ton from. We actually ended up working till around 4am the next day. What I really appreciated is that although it was “an exam” for me, the investment of time on my part was weighed equally by their time investment. Also, I was happy to know that they had learned things during the hack too. I was at ease during the whole day precisely because they treated me as though I was their colleague already.<p>Coming out of that experience, I’ve wondered for a while why employers don’t take the stance that to hire excellence you have to commit time and resources. Most interview processes that I’ve heard of seem to be a cop out on the part of the recuiter&#x2F;interviewer, in the sense that it seems that the priority is to hire the best people but spend the least time doing it. That to me is the ingredient for a broken process.<p>So, in a nutshell, I think talk of take-home interviews misses the point. It might not be scalable, but my intuition tells me that it would be much more effective to have the interviewee and company folks hack on something together. That way, the interviewee gets the opportunity to essentially interview the company too. I’d definitely preceed this with a few conversations on the phone&#x2F;in person, to ensure that the time investment of hacking together is geared towards the best interviewee&#x2F;company relationships.<p>Perhaps this is missing the point too, and perhaps it’s not workable in the real world. My anecdoctal evidence is simply that have spent an entire Saturday working on a project as part of an interview process and ultimately not getting the job, I don’t feel I wasted any time, nor do I feel bitter about it.

I&#x27;ve done quite a few &quot;at home&quot; programming challenges [1][2][3] as well as hacking challenges and most of my experiences have been overwhelmingly negative. The challenge is that &quot;programming&quot; is about 20% writing code and 80% other stuff (requirements, design, testing, security, documentation, meetings, beer, etc...). It seems like the proposed solution is not addressing the problem: Explaining how to solve a programming challenges you may have never heard of on the spot.<p>I do agree that interviews add a level of pressure, but what you&#x27;re trying to gauge thought process. How does the candidate think about solving the problem; Can they iterate; Can we lower the space and&#x2F;or time complexity; Can this be solved recursively, Are they receptive to input? A take home assignment does&#x27;t help you figure any of that out. Usually, the team who is interviewing already knows the answer to the specific problem and all of the intricacies (based on having to solve it on an internal project or just because it&#x27;s a question they&#x27;ve asked before), but the goal is to see someone thinking in real-time.<p>In my career, I&#x27;ve never had a time where I go off on my own, work on something and come back to the team with a solution. Even when I was writing TI-83 programs in High School to solve all of my science homework and finish 50 minute tests in 10 minutes; I still had classmates asking me about what I was doing along the way. There is always discussion and I feel like by sending me off on my own, all of that goes out of the window. You can&#x27;t verify if I copied the code from anywhere else, you don&#x27;t know if I had my buddy who is an engineer at another company break down the answer for me to spit it back to you, and you don&#x27;t know if I paired with someone else to get the assignment complete. I&#x27;m obviously biased based on my experiences doing over 40 interviews in Silicon Valley&#x2F;SF. My goal for this type of interview question would be to understand how the candidate thinks; I don&#x27;t care if you write your answer in Brainf<i></i>k[4].<p>[1] - <a href="https:&#x2F;&#x2F;github.com&#x2F;joeblau&#x2F;sample-elevator-control-system" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;joeblau&#x2F;sample-elevator-control-system</a><p>[2] - <a href="https:&#x2F;&#x2F;github.com&#x2F;joeblau&#x2F;sample-top-ten-tweets" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;joeblau&#x2F;sample-top-ten-tweets</a><p>[3] - <a href="https:&#x2F;&#x2F;github.com&#x2F;joeblau&#x2F;sample-url-shortner" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;joeblau&#x2F;sample-url-shortner</a><p>[4] - <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Brainfuck" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Brainfuck</a>

Software developers huh! What other role requires you to complete an exam to be considered for a job (or even just an interview). Moreover a 4-8 hour exam with no syllabus. No &#x27;past papers&#x27; etc.<p>I think this should stop. In it&#x27;s place, developers can have a pet open-source project, which they submit with their application. The point is that the same project can be submitted for a hundred applications if needed, saving the candidate dozens or hundreds of hours of wasted time.<p>They could put that time into the project instead and even if they don&#x27;t get a single offer, the open source community benefits, and the candidate got to do something interesting rather than convert roman numerals or aggregate an array.<p>I seriously wonder if we should all boycott dev tests. If they give a test, just give them a github link and say &quot;Please review this as a good example of my work.&quot;. If they can&#x27;t review code that isn&#x27;t tied to a silly question maybe they are not worth working for.<p>I say this having done some such tests recently myself. (To be fair one of them said a Github submission is OK, I just didn&#x27;t feel my Github was good enough at the time.). I kind of feel bad contributing to the madness!

So you do the test, but you&#x27;re not guaranteed a position even if your code is amazing.<p>Hell, in my experience you&#x27;re not even guaranteed anyone will read it.<p>This is just a bad idea in an industry full of irresponsible employers. I&#x27;m not saying that Triplebyte is in any way irresponsible, but the rest of the industry has already ruined this avenue.

Only desperate people do these tests. I know this first hand because I&#x27;ve been given a handful of them and can never find the motivation to complete them.

I&#x27;ve wasted too much time on take-home interviews. I&#x27;ll never do one again. All the contracts I&#x27;ve ever gotten never required one.