My friends Andrew and Joel were kind enough to have me back on their podcast Adversarial Learning. We shared our tales of bad data science interviews. Enjoy!
Transcript
[00:00] Tim Hopper: Welcome to Adversarial Learning.
[00:03] Andrew Musselman: Welcome to Adversarial Learning. This is Andrew and Joel, as always, and we have a returning guest, Tim Hopper, who is— since this is his second time, that officially makes him a regular. So Tim, how does that make you feel?
[00:52] Tim Hopper: I’m extremely honored. I’ve never been a regular anywhere except Starbucks when I was in grad school.
[00:57] Andrew Musselman: Okay. So, Joel, you want to introduce this episode?
[01:02] Joel Grus: Yeah. So, this episode is all about interviews, and in particular, bad interviews. So, I have this Medium post that I wrote and never published, and it’s sitting in my drafts folder. In fact, I have a lot of Medium posts that are sitting in my drafts folder because I never published anything on Medium. And it’s basically all of my terrible interview stories, and I have a lot of terrible interview stories. And a number of people have suggested that, you know, this draft has kind of circulated through the back alleys of the internet, and people have read it and—
[01:35] Tim Hopper: 4chan or?
[01:37] Joel Grus: 4chan, 8chan, 12chan, all the chans.
[01:40] Andrew Musselman: Fortran?
[01:41] Joel Grus: No, not Fortran. And they said, you know, don’t publish it, but they’re good stories. I tell them at happy hours, people like them. And so I thought maybe, you know, maybe that’d make a good podcast episode in that they’re out there, but you wouldn’t find them by, say, Googling those particular stories or by Googling my name. And then I know, you know, from Twitter and talking to him that Tim has a number of stories in the same vein. And then Andrew is in the tech industry, so it seems likely that he probably has some too. I don’t know if he does. Maybe all his interviews went really well and he gets every job he applies for.
[02:14] Andrew Musselman: I have a handful.
[02:15] Joel Grus: He has a handful. But what is our purpose here? One, our purpose here is not to name and shame, because we’re not going to name, but mostly just to shame. So that’s one. Two, these stories are kind of sad and funny, and everyone likes stories that are sad and funny. And three, just to let you know that Andrew and Tim and I are all sort of heroes of data science, if you will. And people look up to us and see us as people to be emulated. And so it’s good for you to know that even we, the heroes of data science, go on a lot of really bad interviews and get rejected from a lot of jobs because hiring is broken, you know, and anytime someone doesn’t hire me, that means hiring is broken, obviously. And so you’ll get to hear about how many ways hiring is broken.
[03:03] Tim Hopper: I’ve actually, for those who don’t know, named on Twitter every company that I’ve interviewed at and not gotten an offer from. And it’s linked from my blog post, which is called Some Reflections on Being Turned Down for a Lot of Data Science Jobs, links to those tweets. I decided it’s helpful for people to know that even those of us who have people on the internet who look up to us still have weird experiences and don’t always have success.
[03:32] Andrew Musselman: Yeah, it’s sort of different from the popular imposter syndrome style confessional quotes about, I too am a successful person, but I can’t remember where the backtick key is, or what have you. This is more, you know, just— it’s like we’re all in it together. And it’s hard to interview. It’s hard to be on both sides of the table. So it’s more just like an admission or, you know, an opening of the door, you know, that it happens to everybody.
[04:00] Joel Grus: But mostly to us.
[04:02] Andrew Musselman: Mostly to us because we are bad at interviewing for jobs.
[04:05] Tim Hopper: I’ve actually almost been interviewing continuously for five years. For a variety of reasons, I’ve done, I don’t know, probably no fewer than three interviews a year and sometimes more than that. So I’ve done a lot of these.
[04:18] Andrew Musselman: Well, who wants to kick it off? I mean, I have very few because I’m so good at interviewing, but I could burn one.
[04:27] Joel Grus: Tim’s our guest. Why don’t we let him go first?
[04:28] Andrew Musselman: Sure, sure.
[04:30] Tim Hopper: When I was in grad school, I interviewed at a local technology firm that’s very large and has lots of computers. And because I didn’t have a lot of work experience at the time, they were asking me to talk about my research from grad school, where I was working on stochastic optimization models for scheduling patients in outpatient healthcare clinics. And so these problems that I was working on, I was basically approximating an exact solution that my advisor had tried to do, where he could only schedule four patients before the combinatorial explosion resulted in the model not ever converging. And so I was trying to figure out if I could approximate this problem that was provably extremely hard. So I started explaining it to a guy, and he basically was telling me that this problem I had been working on for six months was a trivial problem, and that essentially just a simple heuristic solution would find the optimal solution. And I don’t know if he was trying to ruffle my feathers or what, but I ended up having a very intense argument with him about how he was completely wrong. I think he had a PhD in computer science.
[05:41] Joel Grus: Who was right, him or you?
[05:43] Tim Hopper: I’m sure I was right, but I didn’t get a job offer, so no one won, I guess.
[05:46] Joel Grus: You won by being right.
[05:49] Tim Hopper: I’m very confident I was right, but it was an extremely hard problem that, as far as I know, no one’s really come up with a good solution for. But I don’t— I just don’t— I wish I knew. Like, sometimes you just wish you knew why someone was asking the question. Like, is he just trying to see how you respond to that situation, or is he really just an idiot.
[06:08] Joel Grus: Or both.
[06:08] Tim Hopper: Or both.
[06:09] Andrew Musselman: Did you get an offer?
[06:10] Tim Hopper: No, I didn’t get an offer.
[06:12] Joel Grus: How’s that company doing today?
[06:13] Tim Hopper: They have roughly 500,000 employees, so they seem to be holding on for now.
[06:21] Andrew Musselman: All right.
[06:22] Joel Grus: But, you know, maybe their time is still limited.
[06:25] Tim Hopper: That’s true.
[06:26] Joel Grus: My first story. The year was 2011. It was sort of, you know, the advent of data science. People were just starting to call themselves data scientists instead of data analysts or whatever it is people called themselves before that. I had a job, but I didn’t like it that much. So this recruiter contacted me and said, “Hey, we want you to interview at this place.” And they basically—here’s what they did. They took TV watcher demographic data and they used it to basically optimize the timing and placement of infomercials by saying, you know what, Fargo, North Dakota, seems like it’s expensive, you know, at the 3 AM time slot, but our demographic data says people in Fargo are much more likely to buy Slap Choppers than people in, you know, Bismarck, South Dakota. So therefore, if you have a Slap Chopper infomercial, you know, Fargo, North Dakota, is a much better buy. So that’s what they did. It sounds kind of dreadful, and I bet it was dreadful, but I interviewed anyway. So I go in and I sit with the guy who’s the hiring manager, and he says, “Okay, here’s what we do.” He explains that to me. And then he pulls up this spreadsheet with like a shit ton of data in it. And he says, “So this is the kind of data we work with. See what you can make of this.” And he hands me the keyboard. And so I’m like, okay. So I start, you know, making summary statistics and pivot tables and crunching through the data and finding some nuggets of information that I can show him to try and impress him or whatever. And this goes on for like maybe 45 minutes and he’s not saying anything. And so after 45 minutes, I sort of, you know, I bold a couple cells and draw the grid and say, “Okay, that’s what I got.” And he looks at me and he says, “Yeah, that’s not how we think about it at all.” And I’m like, oh.
[08:13] Tim Hopper: Was that a criticism or a compliment?
[08:15] Joel Grus: I think it was criticism, but I don’t know. And then he goes, “Here’s how we think about it.” And he takes the keyboard and he shows me like a bunch of other stuff that’s totally different from what I did. And I was like, oh, okay. And so then I left and I called the recruiter and I said, “Wow, that went really poorly.” And he’s like, “Okay, let me check with them and, you know, call you back.” And he called back and he said, “You know what? They really liked you. They just thought you weren’t interested.” I was like, oh, okay. And they’re like, “So they proposed that you come back and meet with their CEO and he can kind of sell you on the vision of the company.” And I was like, all right, fine. You know, I had a job, so I had to like invent another doctor’s appointment now to go back. And so I go and I sit with the CEO. He like shows me slides, he talks through his vision, all this stuff. And I’m like, yeah, sounds pretty good. And he says, “Okay, you know, if you’re interested, our next step is a full day of technical interviews.”
[09:08] Andrew Musselman: Oh God.
[09:09] Joel Grus: And I’m like, what did I already do? Like— and so I went back and I called the recruiter, and I said, “You know, did you know that I was going back to talk with the CEO to see if I could have another day of technical interviews?” And the recruiter’s like, “No, I didn’t know that, actually. This is our first time working with that client. So sorry.” And so anyway, at that point, I withdrew from the process. And I got an email from the hiring manager that said, in essence, “I didn’t want to say this, you know, during the interview process. But now that you’ve withdrawn, like, I looked you up on Google and checked out some of the things you’ve done. And I think you’re an interesting fellow.”
[09:44] Andrew Musselman: Weird.
[09:48] Joel Grus: And that was that. And I don’t think the company— I can’t remember what the company’s called, so it’s hard to figure out if they exist anymore, but I don’t think they exist anymore.
[09:57] Andrew Musselman: Was “interesting fellow” a criticism or a compliment?
[10:01] Joel Grus: I think it—no, I think it was a compliment. Like, I think it was like, you’re a weird, like, Twitter shitlord and stuff, and that’s cool.
[10:09] Andrew Musselman: It’s just a weird thing to email. Like, I always get— I mean, this isn’t a story story, but I get weirded out when I get stalked by somebody I’m gonna be interviewing as a candidate. I mean, I know it’s not weird in reality, but I find it— I just feel like, you know, it’s premature or something. But, you know, I’m kind of a prude as far as that goes.
[10:28] Joel Grus: Yeah, I don’t mind that because I always encourage people when they’re interviewing to find out as much as they can about who’s going to be interviewing them and stalk those people. So it would be a little bit, you know, hypocritical of me to— it is weird when like I’m interviewing someone and they start asking me questions about, you know, my Twitter account, but—
[10:43] Andrew Musselman: Right. Well, I mean, I’ve had LinkedIn requests from somebody who I am about to interview or I just interviewed too. And I find that a little over the line.
[10:52] Tim Hopper: That’s why I always check out people’s profiles in incognito mode first.
[10:56] Joel Grus: Yeah. But sometimes it won’t show them to you. Like, I think they’ve wised up to that.
[11:01] Andrew Musselman: Yep.
[11:02] Joel Grus: I mean, the reality is I’m already connected to like 500 people on LinkedIn that I don’t know. So like, what does it hurt me to connect to some person I’m interviewing? Yeah. Yeah.
[11:11] Andrew Musselman: I have one where I was, you know, I was out of work, which you know, that always puts you in a more desperate position. And the recruiter was the sister of a good friend of mine. And, you know, so she said, “Well, we should get together and talk.” And I said, “Yeah, I have nothing to lose. I need—you know, I need help, obviously.” So we met up and she rattled through some options. And recruiters need to get a sense of where you are in your career and what you’re looking for. And then, you know, their role is to match you up with somebody who’s looking for somebody just like you. And it never works that way. So the recruiter said, she said, “I have something at a company in Factoria.”
[11:48] Joel Grus: They do—
[11:49] Andrew Musselman: So they were—what was this, like 6 years ago or something, or 6 and a half?
[11:54] Joel Grus: For those of you who are not Seattle people, Factoria is like—there’s a mall there basically, but—
[12:00] Andrew Musselman: I was gonna explain how desolate a location it is. But yeah, so she said, “We have a— it’s a company in Factoria.” And for, you know, immediately I was like, “Oh God.” And then, because I, you know, in Seattle, if you have to drive more than 15 minutes, you know, you’re doing it wrong. So, and she said, you know, it’s a social media indexing company. And so what they do is they take all the tweets and all the Facebook posts that they can, you know, buy access to, and they make, you know, they summarize it and they build analytics on top of it. And I thought, okay, I mean, you know, sifting through recycling and garbage, you know, might be a fine way to make a living. And so I went out for the interview, and, you know, I got in the car, got on the highway, found myself in rush hour traffic going out to the Eastside, you know, and it was raining, and it was just the perfect day, you know, just as far as, like, what life would be like taking a job that’s, you know, up to an hour commute each way, depending on traffic. Get to Factoria, I find the place, park, I go in, you know, and you get introduced to people, and they sat me in a room in which I was going to be for seven hours that day. It was not a windowless room. It did have a window. It did have a whiteboard, and it had a table and some chairs. And so we went through, you know, technical questions. We went through questions about my background, which, you know, there was some relevance because I had done some large-scale analytics, but, you know, their specific— they were hiring for somebody who was doing, you know, exactly what they were doing. Again, it’s very, very, you know, targeted role. And the questions got harder and harder the more people they, you know, I got in front of, you know, got to the point where I was standing at a whiteboard with somebody who was asking me the, you know, the question really boiled down to, can you figure out the very tailored, specific solution to this very difficult problem that I’ve been working on for two years? And, you know, at a certain point, I just started shrugging and saying, “No, I don’t know the answer. I can’t— I’m out of ideas.” And so it didn’t go well. And, you know, also, there was a very panicked sense of urgency, like, they, you know, they knew that they were running out of money. And so there was— they were on their last fumes of funding. And so that just, you know, the whole thing felt terrible. And then for lunch, you know, I guess I’m accustomed to, like, if I’m interviewing somebody for even a couple of hours, to offer them to go out to lunch with the team or something like that. And instead, everybody was so busy that they said, “Do you like pizza?” And I said, “Sure, I like pizza.” And so they ordered a pizza and a salad. It got delivered. I sat there by myself for lunchtime eating it, you know, just super, really unhappy and scared. Like, I was already, like, feeling like this isn’t going well. But instead of bailing at lunch, I just thought, you know, while I’m here, I should finish it out. And just, I, it actually just, yeah, it got worse after lunch. So I left, went home, and I was just like, this was a really bad idea. And, you know, the follow-up with the recruiter was just, “No, I don’t think it went well.” And they didn’t choose to move on with the process. So I think it was mutual.
[15:27] Joel Grus: Did you get to keep the leftover pizza? I forget.
[15:35] Andrew Musselman: I forget. But yeah, dodged a bullet on that one.
[15:38] Joel Grus: How’s that company doing today?
[15:40] Andrew Musselman: You know, I don’t know.
[15:42] Tim Hopper: I had a little bit of a similar situation when I also— when I was in grad school interviewing at a company where they put me in a conference room all day, and it was a pretty far walk from my car. I guess that was before I had a smartphone because I was poor, and they misscheduled the interviews. And so I was in the conference room for 30 minutes by myself.
[16:09] Andrew Musselman: Yeah.
[16:09] Tim Hopper: And there was literally nothing. There wasn’t even, like, a magazine I could flip through. No one, like, came to check on me, so I just was sitting there, like, twiddling my thumbs. And I guess there’s that Pascal quote about the problem of humanity with man not being able to sit in an empty room by himself or something, but I exemplified that that day.
[16:30] Andrew Musselman: Well, you made it through.
[16:32] Tim Hopper: Yeah, they hired me, so—
[16:34] Joel Grus: Oh, hey, those sorts of stories aren’t allowed here.
[16:38] Andrew Musselman: Wait a minute.
[16:39] Tim Hopper: Yeah.
[16:40] Andrew Musselman: Oh, am I allowed to talk about interview stories where I was interviewing someone?
[16:44] Tim Hopper: Oh yeah.
[16:45] Joel Grus: I guess. Yeah.
[16:47] Andrew Musselman: Okay, I have one where I was interviewing for a position on a data science team—a candidate who had a PhD in data mining and had a very impressive resume. And so I was excited about it. I thought, well, heck, we can get somebody with some true expertise on the team. And there was a language barrier problem. He had a strong accent, but we worked that through, and we got to where we were understanding each other as far as we could. But I saw data mining and I saw, you know, large-scale analytics on his resume. So I moved into the question, and some questions about matrix math. And that’s when it went weird.
So, like, we had a full hour scheduled. This was about 15 minutes in, and I said, so how are you—how comfortable are you with linear algebra concepts? You know, normally I would expect in an interview like that for somebody to say, oh, you know, it’s been a while, or, ah, pretty good, whatever; what kind of questions do you want to talk about? But instead this guy said, no, that’s too basic. From my perspective, that linear algebra is not a basic set, you know, basic field; it’s pretty hard to grasp stuff in there.
I thought I might have misheard him, and I said, what? And he said, it’s too basic. And I was like, what do you mean it’s too basic? And he said, any high school student could answer questions about linear algebra. And I said, I mean, you know, depending on the high school, sure. And I said, okay, well, I’m not sure I understand what you mean, but I mean, you get the matrices and vectors and dot products and things like that. And he said, yeah, that’s high school level. And I was like, okay, well, that’s fine. So let’s—I mean, we could go up a notch. And I said, so how familiar are you with matrix factorization methods? And he said, too easy. And I was like, dude, you got to be kidding me. What do you mean it’s too easy? Like, that stuff’s super hard to understand. Like, I feel like I struggle with those concepts, you know?
And so I said, I’m serious. Are you seriously telling me that matrix factorization is too easy for you? And he said, yes, this is way beneath me. And I was like, you know, well, look, man, this is the kind of stuff we deal with on the job—rarely, right, but, you know, here and there—and it’s something that we’re looking for. And so I would appreciate it if you’d answer the question.
And he said, look, I’ve interviewed—and he was getting really animated and very angry—and he said, I’ve interviewed with three people already at this company, and now you’re asking me these questions. This is ridiculous. And I was like, you know, I don’t know what to do with this, you know? And so I said, okay, look, let’s move on. Are you familiar with recommender techniques? And he said, yes, of course. And I said, that’s great. So have you—are you familiar with, say, alternating least squares method? And he said, yes. And I said, okay, can you describe how that works for me? And he hung up.
And so, I mean, I was so pleased with that turn of events. I was like, I knew it! You know, like, it was like, I caught you in a lie. But, you know, honestly, I’d hoped that it wouldn’t go this way, and I was very surprised. So I hung on the bridge for another five minutes just in case he had accidentally hung up, and I don’t think he had. So the interview only lasted a half hour.
I gave my feedback to—I was interviewing somebody as a favor because they were like, this guy’s way too advanced for our team to interview. And I was like, I don’t think you want this guy. But that’s one of my favorites.
[20:33] Joel Grus: So I—this was my worst interview for a long time. I got contacted by a recruiter. This is probably about three years ago, maybe four; I can’t remember. Anyway, so the recruiter tells me, you know, there’s this new innovation studio opening up in Pioneer Square, and Pioneer Square is like this filthy neighborhood where all the vagrants live, and then it’s also a cool place to have a startup. And, you know, it’s an innovation studio. They’re going to innovate and spin out startups. And I said, okay, whatever. And he’s like, you know, I think you should talk to them. They have this role. The role was actually called Data Jedi, which should have been, like, my first huge red flag. Because whenever I saw—
[21:10] Tim Hopper: Not Sith?
[21:10] Joel Grus: No, not Sith. It was Data Jedi. And then if you look at their website, all their job titles were like Frontend Ferengi and like stupid shit like that. So I say, sure, here’s my resume. Recruiter emails you back 10 minutes. He’s like, their CEO is super stoked about you. He wants to talk to you on the phone. The CEO was like somewhat, you know, he has his own Wikipedia page. That’s not what he said; that’s what I’m telling you. He was a co-founder of some companies you might have heard of.
So I talked to him and, you know, he lays on all this CEO bullshit about, I believe in, you know, building the best teams, and I believe in hiring people for potential, not what they know. And, you know, one time I interviewed this guy and he was doing terribly in the interviews. And so I said, you know, screw the interview script; tell me about something you’re passionate about. And he was really passionate about baseball or something like that. And so we talked about baseball, and through that I sort of, like, drew him out, and, you know, he was the best guy ever hired. And so, like, that’s like my hiring philosophy. And I was like, sure, whatever.
So then he’s like, yeah, you know, you want to come down to our innovation studio and, like, talk to us? And I said, sure, why not? So, you know, I took a half day off work. Actually, I think I took the whole day off work, and I went down there, and it was very small at this point. It was like maybe five people. And one of them was the office manager, and she greeted me and put me in a room, and she asked me, like, you know, what do you think about our clever job titles? And I said, you know, I gotta be honest with you, like, I don’t wanna be called Data Jedi. Like, I can’t put that on my resume. I can’t tell people that’s what I do. And she looked, like, personally hurt that I had said that, because I think she was probably the one who, like, came up with them or something.
But anyway, so then they bring in this dude who’s like their lead engineer, possibly their only engineer at this point. And he does the same thing that, you know, that one of you guys mentioned earlier, which is, here’s—oh, well, let’s take a step back. You know, this is such a cool building that, like, the conference room doesn’t have any whiteboards. So instead of whiteboard coding, they want me to code on a pad of paper, which is like a real class move there.
[23:07] Andrew Musselman: Yeah.
[23:08] Joel Grus: And so he also pulls the whole, you know, here’s a problem I’ve been working on for the last several months. What can you do with it in 20 minutes?
[23:16] Andrew Musselman: Yeah.
[23:16] Joel Grus: So, you know, I did some stuff, and it wasn’t what he wanted me to do. And, well, I mean, he showed me what he wanted me to do, and what he wanted me to do was like a really clever trick that I never would have thought of in, like, you know, less than several weeks of trying things.
[23:31] Andrew Musselman: Using trees?
[23:32] Joel Grus: Things like that, right? Like, here’s, you know, if we take our data and stick it in this really unusual tree structure, then we can run our models in a certain order that’s like extra efficient, and it’s constant. Well, I mean, so there was nothing, like, conceptually hard about it, but it was like super non-obvious. So no, I can’t imagine anyone would’ve come up with it on the spot. And so I didn’t do well with that. And then he asked me a question, you know, how would you implement something or whatever? And I said, okay. So I, you know, took my piece of paper and I started, like, sketching out, okay, I do this method, this method, this method. And then he, like, stops me and he looks at me, he goes, I just wanted you to say message queue. And I was like, oh, sorry for trying to, like, solve your problem. So he, like, walks off, and I’m sitting there much like Tim was. I had a smartphone, but I was still just, like, sitting there, and I was sitting there for, like, 15 minutes and, like, no one came. And so I, like, walk out of the conference room and I’m, like, looking around the office and I see, like, the whole team is, like, huddled in a corner talking about something. And so I go back and I sit in the conference room some more, and the CEO then walks in and he sits down and he says, hey, you know, we’re a real team here. And one of our team principles is that everyone on an interview loop gets veto power. And the dude you talked to just exercised his. So, like, we’re not going to be going through with the rest of your day of interviews. Then he looks at me.
[24:57] Andrew Musselman: Thank God.
[24:57] Joel Grus: Thank God. This is the best part. And he says, I just want to apologize. Based on your resume, I just assumed you knew something about algorithms and data structures. So, and I’m just like, really? And, like, I was—what’d you say? I was so dumbstruck that I just didn’t say anything. If I’d had my wits more about me, I would’ve told him to fuck himself, basically. And it turns out that, like, several times after that, I’ve gotten contacted by recruiters for that same innovation studio. And every time one of them talked to me, I did tell them to fuck themselves. And then the other thing that’s fun is that one of my Facebook friends is actually Facebook friends with that CEO. So whenever, like, that company, you know, had a GeekWire article about them—GeekWire is, like, the Seattle version of TechCrunch—she would, like, post it on her feed and be like, yay, congratulations, like, CEO guy. And then it would show up on my feed, and I would just be like, you’re— I was like, you’re rubbing this, like, guy in my face. Like, so anyway, yeah, they are, they keep growing. If I go look at their website, they still exist. And, you know, they have, like, probably 50 people on their website now. And they spun out their first product. And I won’t tell you what it is, but it exists, and I haven’t tried it, because it’s not something I would use.
[26:11] Tim Hopper: We have a local firm that I’ve interviewed at actually several times, and they’re—
[26:18] Andrew Musselman: How many?
[26:19] Tim Hopper: Several. At least 3.
[26:21] Joel Grus: Wow.
[26:22] Tim Hopper: The first time in grad school, I tried to convince them to give me an internship that they weren’t, like, planning on. And the guy interviewing me during the interview said, I think I’ve taken up enough of your time, and got up and walked out.
[26:35] Andrew Musselman: Wow.
[26:36] Tim Hopper: And then 2 more times, when I think I’ve been more qualified, I’ve interviewed there and they’ve never made me an offer. But they have recruiters that just recruit incessantly. So whenever I get an email from them, even if it’s anonymous, it’s very obvious what company it is. So I always reply and say, oh, I’ve interviewed at that company several times, so they know enough about me. If they want to make me an offer, you can send it along. But that’s never worked for me.
[26:59] Andrew Musselman: Yeah, I get emails from one of my stories here and there, and every time I’m just, you know, like, no, nope, still no, still no. No, I’ve not changed my mind yet.
[27:13] Tim Hopper: So I have some short stories, but one of my more involved ones is, I interviewed at a company that was based in San Francisco where, before flying to San Francisco, they had me do 4 hours straight of phone interviews, like back-to-back phone interviews.
[27:30] Andrew Musselman: Wow.
[27:30] Tim Hopper: And then they gave me a take-home project where I had 3 hours to complete the project, followed by an hour of doing a presentation to the CEO about my thing. So I did that after work one day. I did a full day’s work, and because of the time delay, I worked on it from 5 to 9, which was very exhausting. And then they brought me out to San Francisco to interview me from 9:30 AM to 5 PM.
[27:56] Andrew Musselman: Well, that’s a good signal that they liked your work enough to fly you across the country.
[28:01] Tim Hopper: Yeah, they actually put me up probably in the nicest hotel I’ve ever stayed at.
[28:04] Andrew Musselman: Which one?
[28:04] Tim Hopper: I don’t remember. I was in a corner room looking down at Moscone Center. It was very lovely.
[28:12] Andrew Musselman: Nice.
[28:12] Tim Hopper: And it accidentally got charged to my card, which is a whole other story. They bring me in to interview 9:30 to 5, which is the longest I’ve ever been asked to come in somewhere.
[28:21] Andrew Musselman: That’s a long one.
[28:22] Tim Hopper: Yeah. So I interview with people all day, and at the end of the day, my last hour and a half is with the head data scientist of this— even that’s a long one. Yeah, it was all very lengthy. And he starts asking me technical questions and then delves into combinatorics problems, just like urn problems, you know, that kind of thing. Which are all fine—like, I love problems like that—but I had just been— I had flown across the country after work the night before, and then I had been interviewing all day. Like, if anything, start my day with combinatorics questions. He starts hammering me with combinatorics questions, and then I had with him one of these similar experiences to what you guys were just talking about, where he was asking me to solve some problem, and I was proposing a solution to the problem, and he had another solution in mind and was very disappointed that I didn’t have— I had my own solution, not his.
[29:19] Joel Grus: Mm-hmm.
[29:20] Tim Hopper: Which seems to be a common theme.
[29:22] Andrew Musselman: It really is. Yeah, it’s a real problem. I mean, it’s like, so when I’m interviewing someone and they’re doing it a different way than I would do it or than I have heard before, I go with it and I’m just, you know, I see where it goes, and if it’s reasonable, it’s fine. It’s not, I mean, it’s, it’s, it’s like it’s not—it’s crazy. It’s crazy how much people own, like, love their own ideas and, you know, how that is such a strong, strong feature of interviewing.
[29:50] Joel Grus: So one time I went to this meetup, and it was like functional programming and drinking. That was the meetup. So it was like six dudes sitting around, like, drinking beer and talking about Haskell. So it was awesome. But anyway, one of the guys at the meetup said, you know, he was spinning out a startup from this organization, and he was looking to hire functional programming people. And so I talked to him, and, you know, we had a phone screen, and we had coffee, and he’s like, “Why don’t you come in for an interview loop?” And I said, “Okay, sure.” So I went into this organization that it was being spun out of so early in the process that I was mostly interviewing with people who were at the organization, not in the startup. And so, like, one of these guys—I don’t wanna give you too much information about him, but he was a very senior guy. He had, like, a really impressive title. And he looks at my resume and he says, “Oh, I see you did social science,” which I did at Caltech. And he says, “Tell me about that.” And I said, “Oh, it’s basically economics, right?” And he looks at me and he’s like, “Oh, I’m so relieved. I was worried it was gonna be like social work or something like that.” And then he launches into this huge, like, diatribe about how, like, he loves economics because, like, profit is the only thing that matters, and, like, making money is, like, so important, and, you know, the economic way of thinking is the only way to approach the world. And I just sort of sat there and nodded for a while. And so I think I did pretty good by that guy. But then that founder functional programming guy was the only—he was the only technical person in the company at that point. So what he did was he had a friend who was visiting in town, was not associated with the company at all, but his friend was just in town. So he’s like, “I’m just gonna have my friend interview you.”
[31:19] Andrew Musselman: Wow.
[31:19] Joel Grus: I was like, okay. So his friend has this, like, little notebook and a red pen, and he’s just sitting there across from me, and he’s like, “Okay, question 1: Can you define what reentrant means?” I’m like, “Nope, I can’t.” He, like, makes a little X, and then he looks at me and he goes to the next page. He goes, “Okay, question 2: Do you know anything about signal processing?” I’m like, “No. Is this a signal processing job?” He’s like, “No, but I like to ask about it. Okay, let’s say you’re sampling a—” So I got that one wrong too.
[31:53] Andrew Musselman: Jesus Christ.
[31:55] Joel Grus: And then he’s like, “Okay, have you heard of this,” and some obscure statistical condition? And I was like, “No, I haven’t heard of that, but it sounds interesting.” He’s like, “How would you test for that?” I said, “Well, okay, well, you know, one thing you could do is you could try and compute this statistic and then try and check—” And so I go off like this for a minute and he stops. He’s like, “No, have you ever heard of the so-and-so test?” And I’m like, “No.” He’s like, “Oh, that’s the test you would use.” And he makes another X. And then he goes on to ask me, he says, “Oh, you know, I see you worked on this in-memory analytics system. Can you tell me about that?” And so I start telling him about it, and then he’s like, “Oh, that doesn’t— that’s a lot less interesting than I expected.”
[32:33] Andrew Musselman: Oh God.
[32:34] Joel Grus: So then he leaves, and then the main guy comes in, and he’s like, “Okay, you know, I just have, like, one question for you. Go to the whiteboard and, like, how would you design Amazon.com?” Yeah. And I was like, uh, okay. So I don’t think I did well on that either. And then, you know, five minutes after that, the guy came back and he’s like, “Yeah, sorry, you’re not senior enough. Have a nice day.” So it was a good day.
[32:58] Tim Hopper: I’ve never been told day-of that I was not a good fit. You seem to have—I’ve been told that at least twice, Joel.
[33:06] Joel Grus: Let’s see. Yeah, definitely. I’m trying to think if there’s another time I was told that. Those might have been the only two times, but yeah, that’s what happens when you give anyone an interviewer veto power, right?
[33:18] Tim Hopper: Yeah, yeah.
[33:20] Andrew Musselman: Well, I like that. I mean, it may be painful, but at least you didn’t have to do the rest of that interview day.
[33:26] Tim Hopper: Yeah, that’s nice.
[33:28] Joel Grus: That one I did the whole interview day. It was just at the end of the day he’s like, “Wait here. I’m going to go figure out what the decision is.”
[33:32] Andrew Musselman: Oh, criminy. Okay.
[33:34] Joel Grus: All right.
[33:34] Tim Hopper: My more common experience is— and this just happened. I just interviewed for a job and was turned down, and they say, “Oh, I’ll get back to you on Monday or Tuesday.” And that usually means one or two weeks. And then they—so you just sit there checking your email incessantly with anticipation of a big raise and a signing bonus, and then they crush your soul.
[33:57] Joel Grus: I have a good story about that, but that’s my last story. So, actually, no, it’s my next story, but someone else can go.
[34:03] Andrew Musselman: Well, I have one. I have one. This was with a local company in Seattle, and it was supposed to be with their recommender team. And so, you know, I went in and got beat up by the recommender team, and that’s good. That’s expected. You would want to make sure you ask the hardest possible things for anybody getting to join your fancy team. And that’s fine. But it turns out—and I didn’t know this at the time—without having informed my recruiter or me, they also interleaved the web services team, which, when I say web services, I mean, you know, the structure of the infrastructure that the entire internet runs on. So probably their most high technical competency team, and they beat my ass all day too. And so, I mean, it was—I could tell something was weird during the interview. I thought, man, this is just really harder than I expected. And I went home, talked to the recruiter, and I said, yeah, it’s kind of weird. And I told her—well, there were other, you know, it wasn’t just the recommender team, it was also so-and-so. And she said, oh, that’s weird. And, you know, I talked to another friend of mine who worked at that company at the time who was a bar raiser, which, if you don’t know what a bar raiser is, that is somebody who doesn’t belong—he or she is not on the team who’s interviewing you, but it’s a cohort of people who are trained at the company to come in and be, you know, at another level and ask the most difficult questions, usually which involve, you know, using a tree structure. I’m not kidding. And so I told him about it and he said, oh, that’s really weird. I have never once heard of—and keep in mind, he’s been on multiple interview loops in his day—I’ve never heard of them putting two teams together on one loop. And I said, oh, well, that’s great. And I didn’t get it. I didn’t get an offer from them. So it was, and that was, you know, it was a really painful day. But yeah, that was a bruiser.
[36:17] Joel Grus: I have a story about the same company. I know a lot of people who work there too. And one of them emailed me one day and he said, hey, Joel, you know, they’re starting up this new team here. And it’s super cool. And I think you’d be interested in it. I think you’d be great for it. And like, I’ll introduce you to the hiring manager. I think you should, like, check it out. And so, you know, this is a friend who did me a solid once, so gotta do him a solid as well. So I went and I talked to the guy who was the hiring manager, and he seemed like a nice guy, and he actually brought a copy of my book for me to sign, which is a good way to kind of, you know, win me over.
[36:49] Andrew Musselman: Nice.
[36:49] Joel Grus: And so I was like, that’s cool. So they’re like, well, you know, why don’t you come in for an interview loop? And I said, sure, okay. So I came in and, you know, the first guy was like, all right, how would you build a spam filter? I said, okay, you know, I know the standard way to build a spam filter, like in the Paul Graham essay. You do a Naive Bayes classifier, you do some smoothing, tokenize the words, and blah, blah, blah. And so I explained that to him. And then he said, okay, now let’s say I’m malicious. And, like, he lists like 30 different malicious attacks against that spam filter. And he’s like, how would you defend against these 30 different malicious attacks? And I was like, I have no idea. And he’s like, okay. And so he just kind of like put me in my place. And then this company is big into behavioral interviews. So everyone on the loop has to ask you behavioral questions. You know, tell me about a time when you, you know, had a conflict with your manager. Tell me about a time when you felt ethical qual—things like that. And literally every single person on my loop said, “I apologize, but we’re required to ask these questions.”
[38:00] Andrew Musselman: Wow.
[38:02] Joel Grus: Every single person on the loop. “I apologize, but we’re required to ask these questions.”
[38:07] Andrew Musselman: That’s awesome.
[38:08] Joel Grus: Tell me about a time when you had to give someone some bad news. It’s like, well, the year was 1980, the place was Strawberry Fields, John Lennon left us all. And I don’t remember that. Yeah, you wouldn’t. And so that was the one guy. The hiring manager came in, he asked me to prove an abstract algebra theorem on the whiteboard. I’m like, does this have anything to do with the job? He’s like, no, but—
[38:43] Andrew Musselman: What does this have to do with putting ping pong balls into a 747?
[38:47] Joel Grus: He’s like, but you have math on your resume. I’ve always wanted to ask someone this question.
[38:52] Andrew Musselman: Get out of here.
[38:54] Tim Hopper: That is so fucked up.
[38:55] Joel Grus: Then his boss comes in, and his boss asked me basically a homework problem about Markov chains.
[39:02] Andrew Musselman: Oh wow.
[39:02] Joel Grus: It’s like if you were studying Markov chains in school and you got to the end of the chapter and it’s like, you know, prove or give a counterexample assertion. And, like, that’s what he wanted me to do on the whiteboard.
[39:13] Andrew Musselman: Wow.
[39:13] Joel Grus: Prove or give it. And so that was that. Then there was another guy who came in and he’s like, all right, I’m here to talk to you about statistics. Explain to me the difference between Bayesian statistics and frequentist statistics and use lots of equations and prove your results. It was like—so it was basically like an oral exam in statistics. So it was like brutal. And then it got worse from there. So anyway, that was over. I was beat. I went home. I think that was like a Tuesday. Maybe on Thursday afternoon I miss a call, and it has a California area code, so I didn’t really miss it. I didn’t answer it because there’s this other recruiter who’s a fucking jackass. Okay, side story: there’s this one recruiter who I have no relationship with whatsoever, but he leaves me phone messages like, “Hey Joel, it’s Dave. Call me back. 415 blah blah blah blah.” Like, he doesn’t say who he is or what he wants, but he’s a recruiter. But he just leaves messages like that: “Hey Joel, it’s Dave calling back. 415…” And so I thought it was that guy, so I didn’t answer it. Then I checked the voicemail and it was the recruiter from this company. So I called back like 30 minutes later, no answer. I called back at the end of the day at like 4:30 PM, no answer.
[40:19] Andrew Musselman: Should have picked up.
[40:19] Joel Grus: I know. Then that was Thursday. Friday, I called at 9:30, no answer. I called at noon, no answer. I emailed my other two contacts, and the message he left says, “Hey, I have an update for you. Call me back.” So I called at noon on Friday, no answer. I emailed the other two contacts I had in the recruiting department at that point, didn’t hear anything back. At like 4:00, I called him one more time, no answer. Finally, Monday morning. So then I have the whole weekend being like, “What the hell is my update? What the hell is my update? What the hell is my update?” Monday morning at like 9:30, he calls me and he’s like, “I apologize, I was out on Friday.” I was like, “Yeah, thanks.” And he’s like, “Well, we looked at the feedback and unfortunately we don’t think you’re a fit for that team. That’s a very special team and they have a very high bar.” He says, “But I also work with 13 other lesser teams. And so if you’d ever be interested in any of those lesser teams that you might be a better match for, feel free to reach out to me at any time.” And I was like, “Yeah, I don’t think so.” So then a week after that, I get an email from the hiring manager—and I don’t think he’s supposed to email me, but he does anyway. And he writes, “Hey, you know, I’m sorry things didn’t work out. I wanted to hire you even if no one else did.”
[41:35] Andrew Musselman: Wow. Oh my God.
[41:41] Joel Grus: And then literally two months after that, I get emailed by a recruiter from that same company that says, “We need people like you.”
[41:50] Andrew Musselman: Yeah. It’s sort of like when you buy a vacuum cleaner and then all your recommendations are for vacuum cleaners.
[41:58] Joel Grus: Yeah, so that was good times.
[42:01] Andrew Musselman: Wow, that’s really good.
[42:03] Tim Hopper: I guess there’s a whole show to be done on the fact that there’s no downside other than a little bit of time wasted for recruiters to send—
[42:10] Andrew Musselman: Oh, it’s in their interest. Yeah, absolutely. They’re not the ones sitting there. They just—all they have to do is, like, fire off emails and then follow up and touch base and circle back.
[42:22] Tim Hopper: So it sounds like you’re a recruiter then, Andrew. That’s what you do all day, right?
[42:25] Andrew Musselman: I love touching base and circling back. I like circling. I was in a meeting one time where—so I did a stint as a large firm consultant. I don’t know, I’ve forgotten already. Yeah, there was a—I was in a meeting where somebody said that we should recircle the wagons without having ever circled the wagons, which bothered me. I have one about interviewing for the recommender team for a well-known movie streaming service.
[42:53] Joel Grus: Go for it.
[42:54] Andrew Musselman: Well, this was before I was a data scientist, but I had built some recommender systems. And so I got an interview with the person heading up the recommender team at this movie streaming service, and I was very excited about it. And I had had good interviews leading up to that. I expected it to be very difficult, and it turned out it was—he was nice, it was great. Didn’t get the offer, but the fun part was when he said, “So how would you build a recommender system for our users?” And I went into the details of, you know, looking at people’s behavior, putting them into vectors and calculating similarity between users, and then looking at, you know, the complement of the intersection of their movie viewing habits and recommending those complements. And so his response was, “Oh, like k-nearest neighbors.” He said, “What you’ve described is called k-nearest neighbors.” And I did not know that at the time. So it was—I didn’t get the job.
[44:01] Joel Grus: Did he say, “Who’s k?”
[44:02] Andrew Musselman: Yeah. No, I mean, so yeah, this is before I had read up on anything, right.
[44:09] Joel Grus: Before my book was available to tell people what all these things were.
[44:13] Andrew Musselman: Exactly. Yeah, long time ago, back in the mists of time.
[44:16] Tim Hopper: Someone referred me to a job at a streaming movie service, but they were uninterested in me not living in a specific part of the country.
[44:25] Joel Grus: Yeah, well, yeah, I mean, you can only stream movies from one place, right? Yeah. Andrew, was it even streaming when you interviewed there, or was it all, like, mail order, or was it a different company?
[44:40] Tim Hopper: Andrew disappeared.
[44:40] Joel Grus: I think we’ve lost Andrew. Good thing I edit these nowadays. Man, I still have one story left. All right, so we seem to have lost Andrew, but we’ll keep going without him for as long as we can. We still got stories to tell, and I think he already told most of his.
[44:54] Tim Hopper: I have a couple just brief ones, brief elements I can share.
[44:58] Joel Grus: Go for it.
[44:59] Tim Hopper: One was I interviewed at a startup here in Durham, North Carolina, where the company was in a house, which was fine. They had this little house out in the country. It was actually kind of cool. But the founder had a large poodle, like a standard poodle, I guess, that stayed in the house. And so while I was interviewing, this poodle kept coming and, like, nuzzling up to me at the interview table. And I’m not allergic to dogs, and I tend to be pretty indifferent towards other people’s pets, but it was very distracting to have this poodle that kept nuzzling up to me. And, you know, what are you supposed to say if someone’s like, “Oh, you don’t mind, do you?” Right? So—
[45:42] Joel Grus: I would say I do mind, but that’s me.
[45:45] Tim Hopper: Yeah, I need to learn to be as blunt as you. And another startup in Durham, North Carolina, the data science team took me out for lunch and then realized I was late for my interview with the Scrum Master. And apparently this Scrum Master was extremely time-concerned, a punctual man, and he would send angry emails to the rest of the company about people being late to meetings.
[46:11] Joel Grus: Isn’t that a requirement for being a Scrum Master?
[46:13] Tim Hopper: So we were walking back from lunch, all the way these guys telling me about how mad the Scrum Master was gonna be. And he tried not to be mad directly at me, but he was really irritated as I interviewed with him. And I had nothing to do with it. Like, I didn’t know what time I was supposed to be back or anything. So that set a bad tone, and they didn’t make me an offer either. Well, actually, I told them I wouldn’t drive 45 minutes every day, and then they said, well, that’s kind of the end of it. So, and I guess to get people to listen to the archives of adversarial learning, they could go back and listen to the time that I had my offer rescinded after I asked about the vacation policy. But I don’t think I need to share that again.
[46:59] Joel Grus: One time I interviewed with a consulting firm for, like, a business intelligence consulting-type job, and one of the people looked on my resume and saw someplace I used to work and said, I know someone who worked there. I will do a back-channel reference. And so they did a back-channel reference on me, but it turns out that that same person who had worked with me at a previous job was also the CEO at my current job. So they called the CEO of my company where I was currently working to do a back-channel reference on me. And yeah, that went about as well as you could expect. That was an awkward day. And it was more awkward by the fact that then the CEO wanted us to have a long conversation about whether I was happy there, but we didn’t actually talk about the fact that someone had told him I was thinking of leaving. We just kind of talked around it, like, what can we do to make you happy here? But without any real context. So it was super awkward. It was awesome. My last big story: I went to—they had this speed dating for developers kind of event. So you’d go there with your resume and you’d spend, like, two minutes at each table talking to a company. And here’s what we’re doing. Here’s who I am, blah, blah, blah. And, you know, most of the companies there were not exciting, but there was one that was a Seattle office of a somewhat trendy startup. And it turned out that someone I knew, a previous coworker, had just joined their Seattle office. So I said to the people there, oh, you must know so-and-so. And they’re like, oh yeah, you know him? That got them very excited about me. It was kind of social proof.
I came in for the interview, and then they just asked me all sorts of stuff that I couldn’t answer. So one of them—I love this question—they said, in your favorite language, write a function that takes a callback and a number of seconds, and then runs that callback after the specified number of seconds. And I said, okay, you know, my favorite interview language is Python, but that’s not a good—it’s not a real good Python problem. Like, it’s not a Pythonic way of thinking. But in JavaScript, we have a function called setTimeout that does exactly that. So I would just use setTimeout. And, you know, I was very proud of myself. Like, you know, I solved it in one line.
And they looked at me and they’re like, okay, how would you write setTimeout? And I was just like, oh, you assholes, right? Like, you’re asking me to—and the job was, like, a big data/machine learning-type data science software-type job. And instead they’re like, you know, we want you to go down to the real guts of, like, you know, threading and concurrency primitives. So I fumbled through it. I did an awful job. I think if you ask me that today again, I would still do an awful job of writing that from scratch. At the end, they’re like, are you familiar with semaphores? And I was like, yeah. They’re like, you should have used a semaphore.
So, you know, they asked me that. They’re like, we have a very popular website. How would you—how would you design it? Or how would you build it from scratch? I was like, oh God. So, you know, here’s your database layer. Here’s your caching layer. Here’s your message, blah, blah, blah, blah, blah, whatever. And so, you know, for whatever reason, I suck at those kind of questions. And then the next guy asked me, like, almost the same question. He’s like, okay, you know, here’s our trendy startup that everyone uses. We want to add this one new feature. How would you implement this one new feature? So, you know, very little of the questions that I do good at; very many of the questions that I do bad at. So it was not a good day. It didn’t go very well.
And so at the end of the day, I’m leaving. I see the guy that I used to work with who’s working there now. It’s his first day in that office. And I go and I say hi to him. And he’s like, you know, how’d it go? And I said, well, you know, it was pretty rough actually, but, you know, I’m sure you went through the same. And he looks at me and he goes, no, actually I worked with these guys before, so they didn’t make me interview.
And that was my last bad interview. I mean, I have a bunch of other stories that are mostly, like, recruiters flaking on me, but those are less interesting.
[50:54] Tim Hopper: I think one of the challenging things with data science—and I guess this is true with other disciplines also—but because data science is such an ill-defined term, it’s really hard to know what someone is going to be talking to you about going in. And I mean, it’s equally hard to know what data science might actually mean at that company. Because one time I was interviewing for basically a fraud detection position at a financial institution in the Pacific Northwest, and so I also flew out after work on a Thursday night to interview on Friday. And I spent the whole flight reading about A/B testing and hypothesis testing, all these statistics things that seemed like they might be valuable. And then they basically just asked me soft questions the whole interview. There was a couple sort of very vague, how might you model this? But nobody had any technical questions at all, really. They didn’t make me an offer, but, you know, that’s typical. But then, you know, so you have that on one hand, and then you have another day you just randomly get combinatorics questions. And how do you even know what to expect?
[52:06] Joel Grus: You just got to be prepared for anything, I guess.
[52:08] Tim Hopper: I’ve just decided that, as the old saying goes, if you’re prepared for nothing, then you’re prepared for anything.
[52:14] Joel Grus: Yeah, yeah, you roll the dice, right? Like, you know, I worked at Google, and I actually had a really good interview there. But then after I was hired, I looked at, like, there’s actually an internal kind of repository of interview questions that you can pull from, but you’re not required to pull from. And I looked through them and saw what the interview questions people used were. And I got, like, a really lucky poker hand in that every one of my recruiters picked a question that I was well suited to answer. I could have just as easily had each of them pick a question that I would have failed, and I would have completely botched the interview, like 0% instead of whatever percent I got. So, you know, a lot of it is just keep doing it. And eventually, you know, I have this oversimplified theory that a software developer or a data scientist is a weighted coin. And, you know, if you’re Jeff Dean or someone, you turn up heads 99% of the time. If you’re me, you turn up heads, I don’t know, 60% of the time. If you’re some of the people I’ve interviewed, you turn up heads 3% of the time. When you go to an interview, they flip you 5 times in a row. If you turn up heads all the times, then you get hired. But then again, I don’t know if everyone has as much trouble with these interviews as we seem to.
[53:21] Tim Hopper: It seems like—I mean, we flipped the coin 3 times and got 3 heads, so it seems like everyone else must be heads also.
[53:29] Joel Grus: Exactly. But if anyone is listening and anyone else has an awesome interview story like this, especially if it fits in 140 characters, you know, tweet it at us and we’ll retweet it, or maybe we’ll read it on the air.
[53:40] Tim Hopper: We can make a Tumblr like that page that used to be big where people would, like, write their confessions on a note card. I don’t remember what it was called.
[53:48] Joel Grus: It was called PostSecret.
[53:50] Tim Hopper: I wanted to call it Postmates, but that’s different. We can make Post Secrets for data science job interviews.
[53:55] Joel Grus: I was actually thinking about this idea the other day. I couldn’t decide whether it should be, like, a Tumblr or a Twitter account or, like, a Hacker News-type site, but where people can just, like, shame recruiters who behave really poorly. So either spamming or writing emails that are grossly inappropriate, and then everyone can share them and cross-reference them and be like, oh my God, that guy did that to me too. I think that could be a fun site anyway. Anyone, anyone have any more stories they’re dying to share? Because we’re hitting an hour here.
[54:23] Tim Hopper: I should interview more because I want to have more stories like you, Joel.
[54:26] Joel Grus: Yeah, I don’t know if I want more stories like me.
[54:29] Tim Hopper: I have a lot of— maybe I’m just not good at making my small frustrations into bigger stories, but, like, I interviewed with a company that asked me to do a homework project, and then, at the interview, they asked me to present the project, which I presented to the CEO. Then when I asked them for feedback as to why they turned me down, they said it was because the CEO was confused during my presentation, but he didn’t ask me any questions the whole time. It’s just like, no. Just, like, really stupid things that happen. I don’t know how you deal with that, but maybe those are good signals that you don’t want to work somewhere.
[55:03] Joel Grus: He wasn’t confused by your presentation. He was confused by the Candy Crush level he was trying to beat on his phone.
[55:10] Tim Hopper: Seems possible. They also put me up in a— I flew to a college town in Michigan, and they put me up in a house that they were renting for their interns. And they—the intern was, like, the host at the home, and he was like, “We just bought some new bedding. It’s in a bag.” This is, like, a bedding set that you, like, you buy for college that was all in a bag on the floor. So I got to make the futon bed myself.
[55:37] Joel Grus: At least, you know, at least you know it was clean.
[55:39] Tim Hopper: Yeah. Yeah. That’s very strange.
[55:43] Joel Grus: Small consolation. I went to—this is not about interviewing at all—but I went to Boston a few weeks ago for a data science conference, Open Data Science Conference, shout-out. And I stayed in an Airbnb because all the hotels in that part of Boston were, like, super expensive. And when I got to my Airbnb and checked in, it had not been cleaned. The bed had been slept in. There was empty Chinese food on the floor, things like that. So I called the host and they didn’t answer. So I left a pretty angry-sounding message. And then I also sent them an email. And then I sat there and puzzled what to do. And 10 minutes later, they called me back. They said, oh—this is one of these things where it’s like a flophouse that someone turned into an Airbnb. So there’s a bunch of units. So they moved me to a different one that was clean, which was fine. But then they didn’t want me to write them a bad review because a bad review is a kiss of death for Airbnb. So they gave me a $50 gift card to a steakhouse that was right by there. And of course, every meal I had was already spoken for except for one lunch. And I had this $50 steakhouse gift card, so I went and I ate a $50 steakhouse lunch. It was awesome. Did you leave a bad review? No, no. I struggled with what to do about that. I ended up leaving no review at all, which I figured was like—I couldn't in good conscience leave a good review because even if I left a good review, it'd be like, yeah, they fucked up, but then they made it right. So I was like, better just to say nothing at all. And then, yeah, it gets lost in the wash. But anyway, yeah, I had French onion soup and then I got this $28 prime rib sandwich, which was phenomenal, and then coffee. And with tax and tip, it was exactly $50. I was very proud of myself. So yeah, anybody who’s listening, if you got good interview stories, let us know because we like good interview stories, as you can tell.
[57:27] Tim Hopper: You should tell people where they can tweet at you.
[57:30] Joel Grus: I usually record that part later, but I do. I usually record that later and splice it in during editing.
[57:37] Tim Hopper: I’m sorry for being helpful.
[57:39] Joel Grus: No, no, it’s good. You’re listening to this, so you found us. But as always, our website is adversariallearning.com. You can go there and find the latest episodes. You can find us on iTunes, or Stitcher, or I don’t know, whatever it is—any of those other podcast sites that I don’t know what they’re called. We love it if you write reviews for us, as long as they’re good reviews. If you write bad reviews, we don’t love it. It makes us sad. And you can follow us on Twitter @Adversarial_L. You can follow Tim on Twitter at—was it TD Hopper?
[58:12] Tim Hopper: That is correct. TD Hopper. TDHopper.com.
[58:16] Joel Grus: TDHopper.com for all your TD Hopper needs. If you want to send us an email, that’s [email protected]. I think either Andrew or I will probably check that at some point. I should probably go check it and see if there are any messages there. I think that’s it. I think that’s our social media presence. Thank you, Tim, for coming on and sharing stories with us and being our first repeat guest, being our first series regular. I’m really— I think this was a cool episode. I’m very excited to get to the editing room and see if Andrew actually appeared in the latter half of the episode, or if there’s just a bunch of awkward silences that I’ll have to edit out. So that will be fun.
[58:52] Tim Hopper: Thank you for having me. Anytime.
[58:54] Joel Grus: Awesome. Good night.
[58:58] Tim Hopper: Bye.