Conducting Technical Interviews, Part I

Generally speaking, job interviews can be awful. They are stressful for the candidate and a huge time sink for the interviewer, often time that has not been accounted for in milestone planning. And the end result of this awfulness is very likely to be an incomplete, perhaps even inaccurate, set of data from which a hiring decision will need to be made.

This week marks my one-year anniversary of working at MobilityWare, and over this last year I have been asked to conduct quite a few interviews of software engineering candidates. I’ve probably conducted interviews at about three times the rate I did them at Carbine. I used to really dread being tasked with an interview, mostly because I suspected that I wasn’t very good at them. And I was right, I used to be terrible. But over the last couple of years as I’ve done more, I’ve started to really enjoy it, and I hope that is because I have gotten better.

Sometimes I interview people alone but often I will be paired with another interviewer, and sometimes we will have three people interviewing a candidate. This has given me an opportunity to not only see how others conduct interviews but also get feedback from peers on my own interviewing. So over the years, and the past year especially, I’ve learned quite a bit. Our HR director and I are doing an in-house presentation on interviewing techniques, and so I’m using these blog posts as kind of a way to organize my thoughts about that.

If you do a web search on “how to conduct a technical interview” you will find all sorts of advice and quite a bit of it will be contradictory. A lot of the writers will be very adamant that their way to conduct interviews is the right way and the way others do interviews is the wrong way. So right off the bat I just want to say that these thoughts about interviewing engineers are what I’ve found work for me. I don’t know if they’re the best ideas and I’m sure that there will be things that you may disagree with. If you are looking for how to conduct a technical interview, then I suggest looking at as many of those search results as possible. If this post also helps you, then great, but the mere fact that you have actively sought out ways to improve your interviewing skills probably means you will be better than 80% of the people out there conducting interviews.

I should say that if you just have time to read one and only one post about technical interviews, it should probably be this one by Joel Spolsky. I mean, I don’t agree with all of it and I think you should keep reading mine, but if you only have time for one, then yeah, follow that link. It’s cool.

Interviewing: Really Important

Interviewing is really important because hiring is really important. Employees are what truly differentiate one company from any other company, and companies that want to do great things must find and hire great people, and it’s really hard to do it well. In my 8 years at Carbine we grew from about 25 people to over 225, and since I started at MobilityWare we’ve almost doubled in size and we are still growing. This has meant a lot of interviewing, and a lot of decisions on whether to hire or, more commonly, not hire candidates.

Don’t Avoid Interviewing

You should want to do interviews. This goes for everyone, but especially for engineers. I know they are a time sink and I know they interfere with whatever tasks you are currently working on, but you should want to be involved in the process of choosing who you will be working with and with whose work you will be interacting. It’s hard to find good engineers and you will often have to be involved in tough “no” decisions, but the consequences of a bad hire are far more costly than missing out on a good candidate. If you have ever worked with engineers who seemed completely unfit for their jobs and you’ve wondered how they got those jobs in the first place, then you should want to make sure it doesn’t happen again.

Even if you are unconvinced that this stuff is really important, I guarantee your HR director thinks it is, so at the very least, you should be happy to be asked to help out with the process. If you haven’t been asked, volunteer. Ask to be part of the process as an observer. Take some initiative and learn how to do this. It’s important. (Have I mentioned it is important?)

Becoming a good interviewer will make you better at doing your job, regardless of what kind of job it is, but especially for engineering. In preparing to conduct an interview you need to be proficient in whatever you are interviewing for. You will need to be prepared to answer questions about the problems you are posing. Despite all that preparation you will sometimes be surprised at a new approach to the problems you discuss in your interview. Remember that many of the engineers you interview will actually be really good at engineering, and so you will actually learn from them as you evaluate them.

The Challenge

So if you are still reading, then hopefully you agree that this stuff is important, but here’s the thing: it’s also really hard. If you conduct the perfect interview, or design the perfect interview process, you will still collect only a fraction of the information necessary to truly evaluate a candidate. You are going to have to make a decision based on incomplete information, so it’s really important that you maximize the time that you have to get as much information as you can to make that decision.

You’ve no doubt heard the expression, “there’s no such thing as a stupid question.” Well, that’s wrong. Or at least it’s wrong if you are interviewing someone. Just as lawyers are taught to never ask a question in a courtroom when you don’t already know the answer, interviewers should never ask a question if the answer won’t help determine whether to hire the candidate or not. You have a finite number of questions; don’t waste them. If I ask the candidate’s favorite baseball team and learn it is the Minnesota Twins, what can I do with that information? (If it’s the Dodgers, sure, that’s a no-hire, but the Twins? I got nothing.)


Prepare a list of questions for yourself about the candidate. These aren’t questions you will be asking the candidate, but questions you hope to answer during the interview. Every question/problem/scenario you pose to the candidate should be about helping you answer these questions. For one of my programming questions, these questions are things like:

  • Did she ask clarifying questions?
  • Did he “define the contract” (API)?
  • Did she get a valid solution? Naive, better, optimal?
  • What was the time/space complexity of her solution(s)?
  • Could he have a discussion about complexity?

Once you have defined the questions you want answered, you need to find questions/problems that will help you answer those questions. I use whiteboard programming questions. Now, some of you just read that last sentence and are preparing your virtual torches and pitchforks. If you look at that list of search results that I linked above, you will find all sorts of very angry writers who have declared that whiteboard questions are dead, that they are worthless, and that they have no place in an interview process. I’ll go into more detail about why I like whiteboard questions, with examples, in a future post, but for now I will just say that I understand why people don’t like them, but I still find them useful.

Make sure to prepare more questions than you will ever need for the time you have been allotted. Do NOT find yourself in an interview with 30 minutes to kill and no questions to ask, or nothing to talk about. I like to have some tough problems ready for candidates who sail through my standard questions and some easier questions for people who just flame out on the standard questions.

In general, don’t try out a brand new interview question on a candidate for the first time. Ask a coworker first, so you can see how it is answered. Your cool new problem might have a really simple solution that you didn’t think of, or it might be too complex for a 45- to 60-minute interview. It might have only one solution, or all the solutions might be of equal time/space complexity, and you probably want to have bad/okay/better/best solutions. Again, I’ll talk about some examples in a later post.

Finally, get together with your fellow interviewers and make sure you have a plan for how you will be tackling the interview(s). Make sure everyone knows what questions they will be asking and when they will be asking them. Don’t find yourself in a situation where you are desperately trying to fill time because all of your questions were asked already. It’s unprofessional and you’re wasting precious time.

The Interview

Now that you’re prepared it’s time to do the actual interview. First and foremost, please try to remember your first job interview. If you’re a normal person you were probably really nervous. Likewise, the person you are interviewing is probably going to be nervous as well. Try to do something to make them more comfortable.

Since I do whiteboard problems, one of the first things I do is write on the whiteboard some information about the first problem. This always involves me needing to draw a set of curly braces { }. I cannot draw curly braces. At all. If I’m really lucky, they end up looking like a capital E and a 3, but more often they end up looking like weird, mismatched squiggles.

After drawing these weird squiggles, I apologize to the candidate and explain that I can’t draw curly braces and ask them to imagine those weird squiggles are curly braces. I do this every time. My fellow interviewers are sick of me doing this. But I’m trying to accomplish a couple things. One, I just want to poke fun at myself to let the candidate know I’m not taking myself too seriously here. Two, I want to “give permission” to the candidate to not worry too much about their own writing. I don’t care at all about penmanship. Or how neatly they draw semi-colons or brackets or even if the brackets are lined up. I’m not always successful with this; I’ve had candidates draw and erase the same line of code several times because they think it’s sloppy and are really trying to make a good impression. I try to encourage them to move on, but sometimes you need to let them do whatever they need to do to cope with the situation. Show some empathy here; it’s tough.

I’ll talk more about the whiteboard problems in Part Two. Regardless of what kind of interview you are conducting there are a couple things you should keep in mind. First, remember that the purpose of this interview is for you to evaluate the technical qualifications of the candidate. You are not there to prove to the candidate how smart you are. I’ve seen this from interviewers that seem to feel the need to show that they belong in the room; trust me, the candidate believes you are a smart person. You don’t need to prove it.

That leads me to the second thing to remember. You are representing your company. Be nice to the candidates! If you end up being impressed with them, you want them to accept a job offer; candidates are much less likely to want to work with you if you demonstrate a thin skin or insecurity in the interview. In fact, you want the candidate to walk away from the interview excited about the possibility of working at your company, even if you ultimately decide (or have already decided) that this person is a no.

After the Interview

Eventually, the interview ends and you have to make a decision. I fill out a score card with notes I’ve taken and I immediately give a thumbs up or down on whether I think we should hire the candidate. There is no maybe. On this particular point, I am in agreement with Joel Spolsky: if the candidate is a maybe, that’s a no. I make this decision before talking with any other interviewers because I don’t want to be influenced. Afterwards, if there is disagreement, and there often is, we will have a meeting where we discuss the candidate. I can be persuaded of course, but I record that initial score immediately.

Finally, and this is something I will talk about in another post at more length, it’s really important to do as much as possible to eliminate bias from your decision. All people have bias of course, but diversity is so important to healthy successful companies. Many companies rate prospective candidates on “culture fit” and that’s okay, but be careful to ensure that you are talking about the company culture, not your personal culture. You want to choose candidates that you can work with, not necessarily candidates that you want to hang out with. Don’t let your company turn into a bunch of carbon copies of yourself; that might sound great to you, but it’s a recipe for disaster.

Of course one way to protect against that is to make sure interviews are not conducted by the same people all the time. Once you have become a regular interviewer at your company start reaching out to other engineers to get them involved in the interview process. Convince them that it is in their best interest to have a say in the process. It will make them better engineers and it will help your company hire better engineers.

Thanks for reading! If you found this helpful or at least hilarious in its wrongness, look for Part Two next week.