Let's break down the data science hiring process!

Charles Givre has personally interviewed dozens of aspiring data science candidates throughout his career. In this episode, he wants to outline his thoughts on making candidates succeed in terms of delivering value to organizations and how to make interviewees stand out.


Lee Ngo 0:04
Hello, everyone, welcome to educative sessions. My name is Leanne Oh, and I am the community manager here with educated, educated makes it easy for authors to provide interactive and adaptive courses for software developers, and educators Sessions is a multi episodic campaign to engage people in the developer world about their coding experiences. And today, my guest is Charles giver, who is the CEO of data distiller. And today we are going to talk about let's break down the data science hiring process, shall we? Charles, welcome to the show.

Charles Givre 0:35
Thank you very much, Lee, thank you for having me.

Lee Ngo 0:37
Charles, let's begin. You know, you're a very experienced data scientist, both as an engineer and as an entrepreneur, and on the management side as well. Let's talk a little bit about your extensive background, especially in building these teams, you know, what do you look for? What are the some of the general qualities that a great proper data scientists? Wow, that's a?

Charles Givre 0:57
That's a great question. I think, and this is a function of a lot of the places where I've worked is that I've been kind of somebody who's been the first person who, the first data scientist to come on board, or the first, you know, the first Trailblazer if you will, in the, in the data science project. So typically, when I'm been building teams, I've looked for people who are comfortable kind of working in those in those kinds of environments where the problem is not clearly defined, where you're going to go in, and you kind of have to tell the either the client or other people how to kind of do their work and not the other way around. And so basically, what I'm getting at is having that kind of entrepreneurial spirit, and also the, the, the acceptance of the fact that not everything will be perfect, I kind of related to, there's the old story they said about American immigrants, when people came to this country, they thought the streets were paved with gold. And then they discovered not only with the streets not paved with gold, they were not paved at all, and they were expected to pave them, I view it kind of the same way with it with data science. And so those are kind of the those to me, are really important skills, especially if you're the first one's on a, in an environment in a company or on a project.

Lee Ngo 2:11
Well, I have never heard that metaphor before. And I like it a lot, as someone who's child of immigrants, and identify a lot with, you know, we bring nothing but our skill set and this unique set of skills that you're going to be very blown away by, I think that's nice and proper. So, um, you know, let's get down into the process of it all. So you've, you know, looked for certain kinds of skills. But you've, you know, part of our interest in this talk is that you've had a lot of different kinds of experiences as someone who's had to interview candidate. So let's talk about the most impressive candidates that you've had, of course, not naming names. But what are some of the things that anecdotally have that you've seen in the past?

Charles Givre 2:50
I think the the most impressive candidates to me, are the ones that just exude a sense of enthusiasm, and love for what they're doing and are really happy to show you what their work. I love personally seeing it when somebody has a project that may have nothing to do with what we're hiring them for, but just how they show how they've applied data science or data analytics to something that they are personally interested in. To me, that just goes so far in convincing me that this person really is going to be a great fit for the team, again, largely because of I've had to work in these environments where you have to have that kind of entrepreneurial drive. But I think it's a great a great characteristic In any event, because it shows that you can apply these techniques to to problems and generate some sort of value from them.

Lee Ngo 3:38
That's great. Are there any other qualities as well? I mean, it sounds like, you know, having a clearly defined passion is great. And I think that's a difficult thing for a lot of people because, especially if they're jumping in discovering that thing that really interests them. What other qualities? Are you looking for?

Charles Givre 3:53
Another big quality that I look for it? Well, two of them, I would say first off is the ability to explain data science, in clear terms to people who aren't data scientists, I think that's a skill that you can't, you can't, that is really, really important for people to have, because many times that your stakeholders will not have an in depth understanding of machine learning. And you it's upon you to kind of explain what is possible, what is not possible how things work, set real expectations. So being able to communicate effectively, I think is very important. Also, along those same lines, I think well, and this is kind of I'm leaving out the fact that I do assume that the person knows what they're doing and has a good solid technical background. So I'm giving that as a an assumed, but, um, I, I think I would just leave it at that is that being able to articulate what you're trying to do clearly to non noisy, non technical, but non data science. People who are not familiar with data science is a very, very important skill. Another thing I would add really quickly is the ability to understand the value that they're creating, and be able to articulate that. This is as distinct from being able to articulate a data science technique. And this is a very important skill as a lot of people, while they want data science in their organizations, they might not really understand why and the value that it can create, and being able to actually articulate that I built you this model, and it will save you X number of dollars a year, or it will generate you x million dollars in revenue, being able to do those kinds of things is very, very important. No, I see

Lee Ngo 5:33
a common trend, a lot of among a lot of the things you've been talking about, you know, his passion, communication, and also being able to, you know, address like issues that are commonly siloed a lot and trying to break out of that a little bit. And, and that's an interesting move, because, you know, I noticed a lot of people who go into data science are very erudite people are speak with a certain kind of a language. But what I appreciated a lot about many data scientists is that they are deliberately working against that they're really trying not to make this as inaccessible. They're actually trying to make this a successful, I should say, and that's a cool thing. So let's get into the less comfortable subject of people that you've interviewed, where it didn't quite go very well. And, you know, let's let's exclude all of those sort of maybe salient categories, like, were they a jerk? Or did they come looking very unprofessional, but more into things that were said or conduct that are generally not appreciated? qualities and what you're looking for? I think,

Charles Givre 6:41
two examples come to mind. The first one was, I interviewed a person, and the person looked great on paper, they had a very, very solid resume. This was for more of a cybersecurity analytic position than a pure data science position. And I usually will start by a softball question just to kind of get the person you know, like, after you've done the intros, I'll throw a softball question out there just to kind of get them talking about their technical expertise. And the softball questions are really meant they're not meant to trip anybody up, they're really just meant to be like, tell me about some really basic technique just to see, and the person had no idea what I was talking about. And that was really awkward, because I had to keep talking to them for about 30 minutes. And that was just a very, very, that was probably that one, I remember is probably one of the most difficult interviews I've ever done. Because I don't like to be rude to people. I'm not going to, you know, just tell them like, get out. You're done. But I thought that was just, that was a very, very special. The other one that that comes to mind was I was interviewing somebody, and the person just exuded arrogance or real condescending attitude. And that, to me is never something that that personally, I don't feel like people who have that those kinds of attitudes are great on teams, because then other team members might not get along with them. So to me, it's important that people are confident, yes. arrogant. arrogant? No, that's there's a line there. But that's always something that doesn't sit well with me.

Lee Ngo 8:21
Gotcha, gotcha. I mean, and so it's interesting, because it's a difficult thing to wrestle, right? You don't want somebody that is, even with us, as simple as questions are, are just, like, in the communication sense, very incompetent, right? And that's something that's going to come up frequently. At the other end, you also don't want somebody that is too arrogant, right? So how do you reconcile people who might not have the skill set or even the communication stylings that you're looking for, but also might have too much of an eagle for you to want to manage? as well?

Charles Givre 8:58
I think what, what I tried to do in an interview context is I really, if someone gets to the point where I'm interviewing them, it's because I've reviewed their resume and want to hire them. So like, at that point, I'm almost kind of trying to bring out the best in them. And so if I get the sense that maybe they're not comfortable talking about something, then I'll try to steer the conversation in a direction that I would hope that they're more comfortable with, but but you're definitely right. I think there is that that kind of balance that you have to have, you have to have that you have to have a degree of confidence that yes, I know what I'm doing. I can, I can explain it without coming across that you're looking down at people and like, I am the all knowing oracle of this and you're just you know, whatever. So there is that fine line of, of, you know, the first of the year, but, but you still that that technical confidence does need to be there and that's there's no, there's no getting around that. Right.

Lee Ngo 9:51
Right. Well, good to know. My last question for you, Charles is if you had to give people three things, they should certainly keep The mind when they're in either a specific data science interview for you, or maybe a more generalized one, what would those three things be?

Charles Givre 10:08
I think first thing is, if you have, make sure you understand what it is that the company is looking for, and nobody's going to be a perfect match. So if you don't have a particular skill, or aren't familiar with a particular tool, be prepared to say, I don't know, tool x, but I know to a y that is very similar to that, and can and a quick learner and will happily do so. So something along those lines, so you understand what it is they're looking for, likewise, understand the interviewer, I have this hypothesis that that interviewers are often looking to hire an idealized version of how they see themselves. So one day, when I retire, maybe I'll do some social research to see if that is, in fact, can be proven or not, but but I think understanding the people you're actually speaking to, is very valuable and understanding like what is important to them, and making sure that you if you really want the position to kind of sell yourself along those lines. And then the third thing I would say is, is make sure that you understand, and I mentioned this earlier, but make sure you understand how to articulate the value of a data science project. Also from beginning to end, too. I feel like too many people get lost in the details of of model tuning and construction and things like that. They forget that the end goal of this is to create some sort of value. And being able to articulate that is very important. Last thing I'll throw in there is there is also the ugly side of data science, which is the data gathering and cleaning, make sure that you can do that and do that. Well. That is such a big part of it. And make sure you're comfortable with that. Because that is Yeah, that's that's like the, the the elephant in the room, if so many data science projects. So make sure you're comfortable with saying I know how to deal with dirty data, if I'm not going to be scared by that, because that's that's often a big problem. So those are my three thoughts. And I see there's a question here.

Lee Ngo 12:04
Yes, I'll get to it. And one more, I do want to note that. Okay, that's a great point you made about data cleaning. And I actually even think it might even be somehow if it can be introduced as an interview question that if you need a data scientist that is 100% resistant to data wrangling data cleaning, in almost a diva like way, you know that they're going to be probably difficult without a context to Yeah, it's like a messy part of the job but unnecessary one and also adept one that demonstrates a skill that's important. So we have a question from Cecilia. The question is, what is your viewpoint on GPA? as an indicator of tech competence? Or in another way? How much weight do you put into GPA?

Charles Givre 12:46
Um, so I guess it would depend on that's the perfect managers consultants answer, there's it depends. I think it would depend on the candidate and how far away they are from school. Like, I don't think anybody would care what my GPA was. But I've been out of school for, you know, I don't want to say how long but let's just say many years. If you're closer to school, then it probably is a good indication of hang on of the person's technical competence, or at least how well they did in class. I think it's kind of just like, for me, it's not a deal breaker. It's, it's just something that if the person was an amazing student, that's good to know, if the person is a horrible student, that's probably good to know. But other than that, I don't see it as a huge indicator, either way.

Lee Ngo 13:37
Okay, fair enough. haros, we've come to the end of our session, believe it or not. And with all of our guests, we like to invite you to look for the shameless plug or a shout out or really any kind of mention you'd like to make to your community. The floor is yours. So

Charles Givre 13:55
thank you very much. Yeah. So I guess the mic. So I recently launched the data distiller, we're a new new startup, we were backed by a very prominent Silicon Valley VC, and we are hiring. So if anybody's interested, if you are a data engineer, or data scientist, please send your resume my way. And I'll happily take a look.

Lee Ngo 14:15
Right? Charles, I'm so great, grateful for you and thankful perhaps maybe we'll make it an topical way. We're being able to speak with me today. And just finally making this happen. Because I've you know, long respected all of the work you've done, and I've you know, I've seen you teach in action. You know, there's just so much more I'd love to pick with you. But for now we'll wrap it right there. So I want to thank Charles for participating in this session. I also want to thank all of you for asking questions and watching as well. Feel free to check more of our talks out on YouTube and also on pod bean as well. And you can also learn a little bit more about us at educated by checking us out@educative.io. So for all of us here at educated Thank you so much. Happy learning. Thank you. Thank you. Thank you for watching this episode of educative sessions. If you liked this episode, please like it and share it with your community to stay informed about the latest sessions subscribe to our podcast at educator sessions dot pod bean comm You can also check out our vid casts on YouTube as well. Lastly, if you're tech curious, check us out@educative.io happy learning everyone

Similar posts