image: thumbnail.jpg
Community Conversation: Hiring Great Data Science Teams
In this panel webinar, you will hear from leaders at RStudio, Boeing, Pandora, Beam Dental, & Amwell, about what it takes to hire a successful data science team:
- Their experiences and perspective on hiring to build great data science teams
- How their teams prioritize skills and experience in new positions
- How they think about skill assessment
- Their advice to data scientists just getting started
As a company that champions data science, we like to make data-driven decisions, specifically around community events like this. If you could take just a few minutes to provide your feedback and what you’d like to see more of, we would really appreciate it: https://forms.gle/zzYMQy4CjUiEedVaA
If you enjoyed this, you might want to attend our Data Science Hangouts. Every Thursday at 12:00 noon ET, current and aspiring data science leaders come together on an open zoom call to casually discuss data science, no registration required. Add the Data Science Hangout to my weekly calendar: https://www.addevent.com/event/Qv9211919
Transcript#
This transcript was generated automatically and may contain errors.
So I'm really excited. We're live on YouTube. If you were here for the Can Do pre-show, you're welcome. That was an unexpected treat for everybody this spooky season, but good morning, good afternoon, and I am so excited to have you all here with us today as we kick off our Hiring Great Data Science Teams panel.
My name is Jessie Mospach, and while I am currently a developer advocate with the Shiny team here at RStudio, not that long ago, I was a data science job seeker who eventually got their first data science job before moving into data science management.
So I know that we have a lot of ground to cover today with our panelists, and I cannot wait for them to have the opportunity to share their advice and experiences with you. So we have four fantastic panelists. Today, we'll be hearing from Rhonda Crate, principal data scientist at Boeing and adjunct professor at Washington State University. We have Katie Schaefer, who is the manager of advanced analytics at Beam Dental. We also have Jairus Singh, director of quantitative analytics at Pandora, and Yiwei Sun, VP of analytics at Amwell. So really an incredible, incredible panel with a ton of advice and experience, and please feel free to put your questions in chat. We do have a list of prepared questions, but as always, we would love to hear from you as well.
Panelist backgrounds and management styles
So to kick things off, all of you come from data science backgrounds, and I'm really curious how that experience has helped inform your ability to manage a data science team. So I'll throw that over to Yiwei first. Sure. So my background is in statistics. That's what I went to graduate school for, and that was a really natural transition. So one of my first jobs out of grad school, the title is senior analyst, but I was working with healthcare data, and the responsibilities and tasks involved R, SQL, working with large databases, doing analyses, basically all of the range of skills you would expect a data scientist to have today. So that experience, of course, was helpful to understand the technical side, but I think just like any other job, whether it's data scientist or, I don't know, marketing associate, anything, the biggest and most important part in terms of getting to where I was is learning how a business functions and just the day-to-day operations, understanding the importance of emotional intelligence, working with other people, being organized, paying attention to process. So those are, I think, more important than the actual data science technical jobs themselves. So it's just really being open and curious, as most, as all data scientists are, to those other aspects of the job.
Rhonda, how about you? I'd love to hear how, you know, coming from a data background has influenced your management style. Sure. My management style is the first question. So I'm not technically a manager. I'm an associate technical fellow, which is a slightly different role than management. We really take on the technical responsibilities of the team in guiding technical vision for execution. But I'll just kind of add to my background is in statistics, and I have a needy cat right now. So I might just hold her here, because otherwise she'll just keep jumping up here. And I have an undergraduate and a graduate degree in anthropology, actually, which is really different than most people in this field. And what it, especially like my anthropology degree, what happens is it just makes me realize the importance of diversity on the team. Because people with different backgrounds and different skills, whether you're from statistics, yes, economics, or, you know, environmental science, everybody brings a unique perspective to solving real problems, right, and how we do that with data. And so really, that's kind of what I look for in a team is, you know, do we have different set of skills? How can we bring those skills and methodologies to solve problems that we have?
Jairus, how about you? Can you tell us a bit about your background and how that helps inform managing a data science team? Yeah, definitely. So I came up through, you know, the team that I now manage. So I started as an analyst on that team and sort of worked my way up to be in charge of it now. And I think the biggest thing is that when you've done the work yourself, it gives you a very, I think, good idea of how to ask people to do that work. And I try to be sort of like the boss, like it, I think it's easy for me to ask myself, like, what was my ideal boss when I was an analyst? And then how do I try to become that? And I think it works well when you've done the work, because that vision tends to be more aligned with something that's achievable and good for the business. But when you're managing people and you don't share those experiences, it's a lot harder to, one, formulate what that good boss looks like and then also to become it. And in some cases, it may be almost unreasonable, right? If you're in charge of a bunch of different groups of teams with sort of competing priorities, you may need to make decisions that benefit one over the other. But having done the work and then having been in charge of the small team, I think I have a very good idea of what it entails and also how to make sure that it's communicated well and it maximizes impact. So sort of being the person I wanted to have when I was doing it is how I go about it. I think it's worked pretty well so far, but that's sort of my strategy.
And I think the biggest thing is that when you've done the work yourself, it gives you a very, I think, good idea of how to ask people to do that work. So sort of being the person I wanted to have when I was doing it is how I go about it.
Cool. And then I guess, yeah, from my perspective, I would kind of echo everything everyone just said. I'd say some ads there would be coming from a PhD program or academia. I think I'm more sensitive on the hiring front to the skill set that they bring, even though oftentimes they will lack an actual job experience outside of academia. Just thinking through their experience as job experience is really important. And I think they can be valuable skill sets to the team, but then also knowing kind of the skill gaps that they'll come with, right, since they have not worked in more of an industry setting. And then I think one thing I didn't appreciate in my old job that I do appreciate in my new job. So my old job, I was a data scientist. There was no really analytics team. It was more of a consulting firm. My new job, I started out managing the analytics team. So I think finally I've gotten kind of a full circle lens of how those two teams can really complement one another, data science and analytics, the differences and kind of the complementary nature between the two. Yeah. So I'd say that definitely informs my hiring of data scientists, knowing, hey, where do we want to take and pick up some of the work that the analysts have done and how will those two teams and two folks work together?
Assessing candidates' technical ability
Yeah, that's really great. And you've kind of cued us up for this next question. So I'm going to start with you, Katie, but I'd really love to hear. I know that everyone watching has questions about this, and I know we all have experience with this, but starting with Katie, how do you feel or what do you feel are the best ways to assess a candidate's technical ability?
Yeah. So a lot of different options for this, of course. I think my goal as a hiring manager is just, again, to your first question, right, having done the work yourself to at least some degree, I think allows you to just assess in some of their answers. Do they know what they're talking about or is there kind of a level of, are they reaching a little bit too far on some of the skills? So I think that my technical experience helps with that, but I think one thing we do here at BEAM is we have a take-home test, simulated data. I think it is really important to assess those technical skills in an interview, but I think the one thing that I'm really not a fan of is kind of, to a point, I think certain more traditional like whiteboarding or brain teaser questions in interviews, I think at a certain point they kind of just measure performance under duress versus performance. So I think there's a happy medium between whether you're doing kind of a live technical interview, live coding, making that productive and worth the time in terms of assessing the skills you want to assess. Take-home is certainly another route that I've become a fan of, but I've seen the kind of more live technical interviews done well and then some maybe not so well where, you know, folks might have missed out on some good candidates just because they asked a brain teaser that the person didn't get in those five minutes. So that's kind of my quick take.
I'll kind of follow up. So I'm a huge fan of them basically talking about what they're passionate about and describing situations where they've actually applied it, and some of the key criteria, you know, that we probably look for in some of those answers are like, you know, is somebody coming up to you and asking you for advice on how to solve these problems, right? Are you being a mentor to somebody else? Because that's usually, you know, a way that you can share your skills with somebody else. That means you're confident enough to to demonstrate your skills and then share that with somebody else and basically teach, right? Also your, you know, your ability to describe it in a technical manner in a situation which you applied it because honestly, like, you know, people come from a variety of different backgrounds. There's a whole bunch of different language that different backgrounds use to describe the same processes. And so, you know, I really like the approach of asking them to describe a situation that they've been in, what was the outcome of that, and really describe the result in a technical format. And you can usually tell when somebody's passionate about something and let them kind of pursue their passions a little bit. But at the same time, making sure that the skills that we need for the job, right?
I agree. Technical exercise is really important. I've always used them. It's, you know, it's it's really the only way you can quickly and accurately assess the technical skills. So I think it's required. Also, just, you know, the basic integrating and job seeking 101 expectations, you know, well put together resume, of course, you know, no typos, things like that, things they teach you over and over again. And then over time, at least I like to think I've developed like a intuition also, you know, it's well, it's more than intuition. It's on the resume. It's also in the actual interview process themselves. You can start to assess to the extent, you know, that person's really curious and passionate about the work itself.
So curiosity and energy to me are my number one and two traits that I look for. You know, if you've only learned how to write R on the, you know, for a couple of years or in grad school, I think you can pick those things up, but you can't you can't teach someone to be really curious and really energetic, you know. Yeah. And then finally, when we do start having technical discussions, I think this is reflected in what Rhonda said, you know, it's a matter of does this person know what they're talking about? And so I listen for are they being specific in what they're describing? If they start speaking in generalities and, you know, vague answers, you know, pretty quickly, they don't really, they aren't really well suited for this kind of job.
So curiosity and energy to me are my number one and two traits that I look for. I think you can pick those things up, but you can't you can't teach someone to be really curious and really energetic, you know.
For sure, yeah. I'm going to echo a lot of things that were said. I'm on team take home as well, I hate doing brain teasers in interviews or even asking people to do them. It's, I don't know, it feels like torture a lot of the time. But a plea for hiring managers out there, you know, try to make sure that you're asking sort of as little of folks that you can to get a good assessment on them to just be respectful of their time. Because what happens is, and I think it works in your favor, because you have some very bright folks out there who are like, hey, you know, I don't know if I need to invest four hours into this when I have two offers. So you're going to turn away folks who, you know, really do know what they're doing, if you make it too sort of heavy. And the opposite logic, which I don't agree with as well, if they really cared, they would sink in all this time and energy. And I just don't agree with that. But for us, you know, I think there's a big concern that, well, what if they get a lot of help on the take home, you're not really assessing their ability. So we like to ask a lot of follow up questions during the interviews. And we usually use one of our interviews to say, okay, you made this decision here, you made this choice, or using this technique, can you tell me a little more about it? And then we're doing what you said, you know, is this something that your friend told you to do, and you have no idea how the technique works, and you can't answer questions? Or is it something that you have knowledge of? So that's how we sort of try to bring that all together in our interview process.
Adjusting assessment based on educational background
Yeah, that's really interesting. A lot of great, great perspectives here. And we've had a question from the audience that really dovetails nicely with this. So I'd be really curious, if you change the way that you assess a candidate based on their kind of educational background. So for example, you know, data science degrees are now a thing that people can get. But if someone comes from a quantitative background, computer science, statistics, data science, versus someone coming from, you know, just a non quantitative background, education, or, you know, policy or literature, do you change how you assess that individual in an interview? Or, or no, maybe everybody's a no, but I'd be really curious if there's anything, how you kind of may approach that differently, if at all.
I'd say that the first thing that sticks out to me is a little bit less of what I do kind of where the conversation heads, but to me, it's really important you, whatever your applicant management system is, it's really important that you're keeping up with the recruiting, like the lead recruiter on it to make sure that person gets an interview. That's what I found to be the most crucial way in which I step in, which I typically wouldn't is just giving the head recruiter that like, commenting when we see the application and staying stupor on top of it myself, because it's a little bit harder sometimes for them to spot like, Oh, well, this person with this non traditional background has the skills you want or not. So I just make sure that I get a chance to talk to that person. If I see that they've got some components that are maybe non traditional, but speak to their skill set.
Yeah, and I'll kind of add that as somebody who comes from a more non traditional background, right? I graduated with a liberal arts degree,
and then I got my stat degree. But I will say like, personally, as I entered job interviewing, I was worried that my anthropology degree would actually set me back. But the reality, I mean, still had a stat, like I still have a master's in stats. But the reality is, it brought an element that differentiated me from the candidates. And I will say, at Boeing, everybody gets the same opportunity of interview for that step. So everybody gets the same question. So as far as that goes, that that's a standardized process that our HR has developed, right.
But I would encourage like any people that just have a solid, like science background to go out and discover the liberal arts, because it can only make you better, right? That's my personal view.
I love that, right. So if you're coming from a quantitative background, go out, discover, you know, do things outside of your educational background. So this really ties into a question, and I'd love to start with you, Jairus. Skills like empathy and communication are often mentioned as like absolute must-haves for data scientists. But how do you know a candidate has these skills? Yeah, I went through this great training at work a while ago. And it was sort of a generalist training. It was not for hiring technical roles. It was just hiring in general. And they went over a really good framework for asking behavioral questions. And what I thought was so unusual about it was that like the questions to me seem to have obvious answers.
But if you ask 10 people or 10 data scientists, you're going to get rid of three of them because they're not going to give you the answer that that you would want. And so what we do is we sort of ask these, we put them in situations that are sort of undesirable. And I don't mean literally in the interview. I mean, we ask them about situations that are undesirable. You know, you've gotten a vague request from a stakeholder, and we see how they deal with those situations. And I think folks who deal well with that, you know, first say, okay, you know, I get that this happens on the job, and I will work with them to come to an answer. But there's always a subset of folks who kind of put it back on you and is like, well, isn't your job to prevent these things from happening? Or like, why would I be in this position? And it's, you know, as much as I'd love to live in a world in which it never happens, it always happens. And so it's kind of an easy way I've found to ask those behavioral questions about, you know, difficult situations that involve working with people or communicating things to people, and seeing sort of how flexible data scientist applicants can be in their answers.
So it was funny to me, because when I first learned these, I'm like, well, no one's going to say, like, I shouldn't do that, or that's not my job. But you end up picking up a few people who genuinely do feel that way. And, you know, I'm sure they're a great fit for another role out there. But that role isn't one of ours.
I'm sure they're a great fit for another role out there. But that role isn't one of ours.
Knowing when to grow your data science team
Yeah, for sure. And, you know, it's kind of nice to get to that point where, you know, right, like, you don't want to pull a candidate along if they're not a good fit. So as we're talking about hiring, I'd love to hear from you to kick things off in terms of how do you know when it's time to grow your data science team? And kind of to piggyback off of that, what strategies do you employ when you're growing your data science team? And what strategies do you employ when you're growing your team?
Everywhere I've been, it's always, well, I've always been hired because they needed to grow the team. So they hired me to grow the team. So it's, it's all the answer is always pretty much most, you know, most for most firms that are data rich, obviously could stand to use more data scientists. But, you know, to answer your question directly, it's really a matter of demand, right? Last I've been doing this for a long time, and we just never have enough data scientists.
The type of teams that I've managed have always been internal support and external support, supporting clients and supporting other functions like marketing, sales, operations, etc. So the appetite for data is, you know, bottomless. So we're always in growth mode. And then when it, when it comes to actually building out the team, which I'm currently in the situation, you know, I really want to try to build out a team that's diverse in all aspects, you know, technically, and experience wise, you know, there is no such thing as a data scientist who can do everything. I also don't want to tame a data scientist where everyone has the exact same set of skills. So you're gonna have people who are more experienced with, you know, on the data engineering side, people who are more experienced on the statistical modeling side, etc.
In terms of experience, you know, having people from different backgrounds, in terms of degrees, that's a great thing, because they bring different skill sets and people from different industries, you know, because that allows, you know, everyone brings their best practices, and we can learn and teach from each other, learn and grow from each other. And that's, that's how I've, I've done that. That's how I've done it. And it's worked well so far.
Yeah, I really love that statement that everybody brings their best practices. I think that that's just such a wonderful way to think about about growing a team. So I would also really love to hear from folks. We have a question from the audience around managers staying in the code. So in other words, what are your thoughts and feelings about management and being a manager? And should a manager still be able to write and review code as well as solve technical questions? If you're at a startup, yes.
Managers and coding
Yeah, I've, I've flamed out a few times at a few startups, and you know, not surprisingly, you're, you're short staffed there. So you're expected to pitch in. But in my experience, and this is just my experience only in healthcare, you know, once you go beyond the director level, I think it's important for the manager to actually be the, you know, the one well managing, because again, this goes back to my firm belief that process is really important. You know, you can't just have people just running around doing whatever they want to do. You have to steer the ship, everyone in the right direction. It's a lot of work, you know. Get, I can still read R, you know, a little bit, but you know, I won't be nearly as efficient as the people who are actually as the people who are actually much better at it. So you don't want me writing the code.
You know, instead, you want me going out there, you want your manager going out there and, and thinking of it this way, finding more business for the team, because finding more business for the team, whether it's internal or external means more opportunities for you to do other things, you know, work on other subject areas, add additional technical skills. So that's, that's how I think, that's how I view myself.
I'll kind of add, because I'm not a manager, right. But that's one of the great and awesome things about Boeing is we kind of have different career paths. So you can go into management, or you can be just a really awesome employee, or you can go into the technical fellowship, which is a career path driven towards technical excellence. And like, like, I'm an R Boeing designated expert, right? There's not very many of those at the Boeing company. And, and so that's awesome. Like, I can help you program and management can really focus on those more HR and going out and getting the business thing. While at the same time, technical fellows also can kind of participate in that going out and getting the business and where the technical vision and things like that are. And so I think I like Boeing for that, because if you really want to focus on technical excellence, you can really drive your career in that direction. You don't necessarily have to end up in, in management.
Yeah, that's great. With that being said, though, managers should still be data literate. I love it. I love the enthusiasm. Katie and Jeris, anything to add on that? I'm seeing a lot of nodding. So I'm not sure if we're all like, yeah, this, I agree with this. Or maybe someone is like, no, absolutely, managers should be coding every day. I think it really, just like others have said, depends on the stage of the company. So I think at a certain point, it while it sounds nice, I think it'd be fun. It can be unrealistic to expect that somebody who's managing multiple different teams or folks has the time to program considerably every day. But I do think you need to have a level of competency. And ideally, you have had experience in the role or akin to the role in your history. Because I think if you don't have enough experience, it can be hard to spot the root cause of some like personnel issues, or maybe other friction you're experiencing. Because if you can't go in and assess the issues with the code or the performance, it'll be hard for you as a manager to do that.
But at a certain point, hopefully you staffed up a successor who is able to do that for you. So you can focus on the other kind of higher level things in terms of where are we taking the team? What kind of projects are we taking on? But for me, it is important that I do write code from time to time, especially as projects are kicking off. But certainly, I think it depends on where you're at in terms of whether that should be a daily thing, or you should be actually moving projects forward yourself. But I think having that skill set is important to make sure your team's functioning properly, and you're hiring the right people.
Hiring for potential to learn and grow
Yeah, absolutely. And I love this emphasis on, you know, you've mentioned where you, you know, if you're starting a project, you're much more involved with code. And Yiwei mentioned a lot about like the stage of the organization. So I kind of want to pivot a little bit, and I'll throw it to Jairus first. But I want to talk about something that, at least on social media, I think gets a lot of play. And that is about hiring someone with the potential to learn and grow. So we see a lot of advice from things like, I don't meet 100% of the requirements, or I only meet 60%. And so you'll see mixed advice from things like, oh, as long as you meet 50%, you should apply, or you know, it doesn't even matter, just apply for it, like, just go for it. And then you'll see people were like, no, you've got to be able to do 80% or 90%. So when we're thinking about hiring someone with the potential to learn and grow, how do you assess this in a candidate?
Yeah, that's a great, that's a great question. I think almost everybody is in this boat where they're not matching all of the, you know, criteria that have been laid out. So, you know, if the company does indeed, you know, box number seven does indeed need that, like you have to convince them that this is something I can pick up over time. I think one of the best indicators is just a track record of having done this in the past. So when you're positioning yourself for the role, if you can look at prior experience and say, you know, I did this role, and I didn't have XYZ, but these are things that I picked up in the context of it. I think that's a really strong, you know, you're showing that I've done this before, and I'm looking to do it again. So that's the best way I've seen it positioned, because you do have a lot of candidates who say, yeah, I'll pick it up, but only a subset of them are like, here's sort of my track record of having done this in the past in prior roles or in school or, or what have you.
Yeah, I think actually school is a good example as well, because you have folks who need to learn a programming language to do a specific, you know, project. And so they can, they can bring that to light as well. But I think showing that track record is really effective, because it goes beyond saying, you know, trust me, I'll do it. But like, trust me, I'll do it because I've done it before.
I think showing that track record is really effective, because it goes beyond saying, you know, trust me, I'll do it. But like, trust me, I'll do it because I've done it before.
I would add to like, if the candidate expressed any interest in continuing education programs at the company, right? So Boeing offers a lot of continuing education programs. And so if you're like, really interested in learning and expanding one area, say like, oh, I'm really interested in geospatial data science, right? Because there's not a lot of opportunities for that, maybe when you went to school or whatnot. And maybe that was one of the small things on the resume, and you're like, hey, do you offer any internal programs towards this or somewhere where I could learn and expand that particular skill? I always look for those kind of opportunities in the area too.
I'd love to hear from other panelists. I think my challenge is doing this for more junior candidates, where often there isn't that experience. So you kind of have to go off of how they answer things. If anybody's found a magic bullet there, I'd love to hear from it. But I think that can be the most challenging spot for me is when somebody is junior, you don't necessarily have the expectation that they would have had that experience because you're hiring a junior role. To me, those are the biggest blind spots, other than kind of how they approach the project, how they answer questions. You have kind of the least amount to go off of to know, okay, I know this person is going to need a lot of support when they start, but how do I get more reassurance that six months from their starting point or nine months from their starting point, they're going to be one of those people who's just kind of self-sufficient. So kind of an ask comes in, maybe it's more vague. They will develop that kind of more harder-to-find skill set, harder-to-find skill set of managing and kind of doing things from start to finish on their own. I think that can be a little more challenging for the junior folks.
Go ahead, Jairus, please. I just had a real quick thing I saw someone do during an interview that stood out, which is when you talk about concepts or techniques that the person is not familiar with, people with that sort of growth mindset will make a note and engage with you more and say, I want to learn more about this later. Some other folks, they get really embarrassed, I don't know that, and they try to steer the conversation away. And I thought it was impressive when someone both admits to me, I hadn't heard of that, and I will go look it up later. And then sometimes you get the follow-up email, hey, I read about this. So that really stood out to me as a positive behavior that I think indicated the potential for growth in the future and the interest as well. I'll throw it to you, Ewai.
Sure, thanks. A couple things. I think everyone who applies to be a data scientist at least has one thing they're more comfortable with. So whether it's, again, statistics or programming in one language versus another, I ask them about that and how they came about to pick up those skills. It's also, you can see it also on their resumes and in their conversations too. Like, hey, I stumbled upon this because I was a graduate student in this field and I started doing data analysis and I realized this is what I want to do, not
ex-engineering. I don't want to pick on a particular engineering, but somebody in an ex-engineering field does a lot of data analysis and realize their passion is in data. And so tell me what specifically about it. Do you like the programming part? Everyone has a thing that they're pretty good at, but in the other aspects of data science.
And or if they're much more junior, I ask about their interests outside of data science. What's the thing that you love doing? Is it ballet, dance, or cooking? There's always disciplines that transfer really well because in my mind, it's about process and practice. I don't see data science as any different. It's process and practice.
Yeah, I will just add kind of quickly from my experience as being an adjunct professor, and it is interesting moving from like this graduate frame of mind to the fact that, you know, certain universities are now offering undergraduate degrees in data analytics, and the breadth and experience is a lot smaller. It's a lot smaller of a window, right? And teaching a senior capstone class is really giving me insight on how students actually think about these and then thinking about how does that translate when they go and try to interview for a job? Because, you know, obviously, like most of them probably just went through the general education requirements. And then those last two or three semesters of school, they were really cramming in like all these kind of analytics classes. They don't know how to describe a logistic regression to you. You know, they're like, they're just, you know, really starting at that very fundamental phrase. And I will say like, I enjoy also mentoring the students because in the capstone class, we spend some time talking about how do you do interviewing strategy, especially like, on your resume, you were, you know, a, like you worked at the deli at Safeway, or something, right? And so I talk about, really expand on that experience working at the deli counter, you know, talk about, like, you were employee of the month, why were you employee of the month, right? Really accentuate those, those skills and talk about your areas and passion of why you picked this major to begin with, and then how can you apply those to this job if you're given, you know, the opportunity to do that. And that's, that's usually just general advice I give the students, because, yeah, it has been enlightening seeing it from both ends.
Weighing candidates' non-technical contributions
So, yeah, absolutely. And I think something that everyone's touched on is this idea of being able to articulate what skills you have and how they translate to data science. So not all of your skills have to be in data science, but they do still relate to your ability to do data science. And related to that, we have a question about thinking about candidates, and how, if at all, you weigh their non-technical contributions to the broader data science community, whether that's through social media, blogging, YouTube tutorials, anything that a candidate does, that's maybe kind of in this amorphous, they're not doing it for work, they're doing it because they're really curious and passionate. Does that come into play for you when analyzing or kind of deciding on a candidate? And if so, how?
Not for me. I mean, it's, to me, their social media interactions are, you know, things they do on their own time, which is great, but I don't think it has any bearings on the actual application. So, let's see, do I think they can do the job or not?
I'd say I have a slightly different perspective. I'd say, at the end of the day, my focus is definitely, can they do the job or not? So it's not a requirement by any means. I feel like I've hired great folks who don't have some sort of blog or data science side hustle, but it is a plus when I see that, because to me, I think what I do care about is that you are, depending on the level, if you're a more senior candidate, that you are going to be staying on top of, like, what is new, new packages, new trends, new programs, new methods, right? Like, if you're, if I'm hiring you for doing NLP analysis, like that space is changing so constantly, it would be impossible to think that one, you know, person could keep up with that and all the different things, right? So if that's what you're going to be a specialist in, I think it's a great, like, plus to show your hunger for that. But I think I've definitely seen folks who demonstrate that hunger and that kind of up to speed with current state who maybe don't have a blog or don't run a chapter of, you know, PyLadies or RLadies or something. So I think it's a big plus, but I think you can demonstrate it in other ways. So that's how I view it.
Yeah, and I'll just add, like, you know, at Boeing, we're not necessarily allowed to participate in certain things. And so I always just want to add, like, as other hiring people, sometimes that should weigh in as part of the consideration as well. But with that being said, especially, like, if you're working in a research and technology branch, I would expect you to participate in professional associations, right? Especially, like, as part of, like, the technical fellowship, some of it is, like, your external connections and how people are looking at you technically. But that is, like, at a level of technical excellence, right? Versus, like, I just like to write and blog. So I would say, like, during our interview, I can't really give you an answer to how that weighs at Boeing. I mean, to say that sometimes consider that maybe those people are not actually allowed to participate in those channels as well.
Yeah, to add to this, this is a really tricky one for me, because as much as I think it's fair to say, yeah, the person's showing passion, they're showing the ability to maybe build communities or to engage, you know, in a broader sense in data science, I don't want to create a world in which future data scientists feel like they need to have a blog or they need to have an active social media, because some people have no interest in doing that. And I think they should be able to get data science positions without having to do it. So I haven't had a lot of candidates cross my desk that do have, hey, check out, like, my data science Twitter. Personally, I enjoy them. But I would just hate to pick someone over someone else, because, you know, someone is sort of more naturally inclined to do that sort of work than someone else. And then, of course, it depends on the quality of the work. I think something as simple as, like, I have a friend who works for a large tech company in advertising, and every week explains an advertising term to all of his followers, many of whom don't work in advertising. And I thought that was a really cool, simple way of teaching folks more about sort of the ecosystem that he works with. He's in sales, so it probably helps him generate more. But just a really small, elegant way to show, like, I have knowledge and I want to share it with the community and it doesn't seem like too much of a heavy lift. I don't think he's applying, but if he was, I feel like that's something he could bring up. And maybe that's a lighter weight way of demonstrating that you have the ability to communicate and participate.
I don't want to create a world in which future data scientists feel like they need to have a blog or they need to have an active social media, because some people have no interest in doing that. And I think they should be able to get data science positions without having to do it.
Yeah. Go ahead. No, no, no. Go ahead.
I'm really glad this topic is being discussed. I agree with Katie in that, you know, it's important for people to stay up to date. You know, I love Twitter because, you know, I'm an avid follower of folks on, you know, Stats Twitter, R Twitter, Machine Learning Twitter, whatever. You know, so that helps me stay up to date and find resources there for sure.
So this is more of a cautionary point in that I also see a lot of data science enthusiasts on Twitter who, I don't know if this is true or not, I worry that they're learning data science or stats too much from Twitter. There's only so many things you can, there's only so much you can learn from, was it 280 characters? I can't remember, you know, and you really have, you know, this, this is a slow skill that you have to acquire with hard work and not, not a fast skill that you can just pick up with tweets. You know, Twitter is great for finding links and, you know, to longer blogs or resources, but, you know, I would not substitute, hey, I, I followed this 20 tweet thread and now I know, you know, X skills, right? That's just not going to happen.
Yeah, I think these are all fantastic. This is, there's a lot of really great advice in here. And I think social media is very rewarding, but it's also like that quick kind of reward. And I love this idea of thinking about, you know, using social media as a starting place or as an inspiration, but to make sure you're, you're kind of doing, don't let the underlying work of learning to be a data scientist and continue growing your data science skillsets suffer for your social media. This is very hard for me to say as someone who lives on social media, but yes, do the work. Social media is like the candy later.
Data analyst vs. data scientist
So I do, there's one question we're going to kind of drastically shift gears a little bit, because I really, really, really want to make sure that we get a chance to answer this. And Katie, I would love to start with you. I would love to know how you think about and define the differences between a data analyst and a data scientist.
Yeah. So I'm sure this looks different, a lot of places, but one kind of division of skills that's resonated for me. So from a technical perspective, if I think of where like our analysts versus data scientists are spending their time, I think a lot of analysts are spending and tend to be a little bit stronger on the SQL. Typically, just to be honest, typically, just to be honest in the industry, BI tools, so Tableau, Looker, Power BI, and then my ideal analyst would also be able to dabble in Python or R. So they'd be able to pick up R or Python for some deeper analyses when they need to. I've come to realize when I got out of consulting though, that a lot of times they're super strong in SQL. They do a lot of that in SQL, but we found opportunities for doing something in R or Python instead when that's the better tool for the job. And I'd say from a job perspective, analysts tend to be, just have a greater variety of projects going on that tend to be more focused around business problems and kind of maybe more qualitative post hoc analyses versus forward looking models. To contrast with the data scientist, R and Python, like strong expertise. And one of those is a must versus kind of a nice to have, or I will teach it as we go. The background statistics is also crucial. I think some kind of foundations are crucial for an analyst, but I wouldn't necessarily expect an analyst to be extremely deep. Just in the way that I wouldn't expect a data scientist to maybe be apt for sitting in on a bunch of business meetings across a variety of areas and answering a variety of one-off questions. The data scientist time should be more focused on protected on specific projects. And so it's to me from a skillset perspective, that statistics training, R, Python, high level of competency that kind of set things apart. And a little slightly less kind of forward facing with multiple different kind of stakeholders from a business perspective.
So I will say my first job was a data analyst because there was no title data scientist. And I did the same exact thing, in all honesty. But with that being said, at the Boeing company, it would probably come up as a programmer analyst job, kind of what Katie was describing. And so I think it's really important to not necessarily pay attention to the title, but really read what the job description is saying and ask questions when you get an interview about the job itself. Because yeah, it's just, there's so much variety. And like you could be a research analyst at Amazon or something, for example, right? So.
Yeah, if I can chime in here, like, I don't know who's asking the question, they might be a job seeker. But, you know, I can give my definitions. But unfortunately, when a lot of people in the industry don't adhere to, you know, a strict definition of one or the other, I think the most important thing to do when looking for jobs is, as Rhonda said, what does the job entail? And not just what's written, but ask follow-ups, because sometimes they say you're going to be doing seven things, but it's really two of those seven things that you're going to be doing. And so try to narrow in on that sort of as gracefully as possible. But the thing I've seen happen in the market, because I mentor students who are going on to get their first jobs, is that, you know, data scientists, a lot of people put that above data analysts, and certainly in pay, that happens a lot. But you have a lot of companies now treating, you know, a data analyst role and calling a data scientist to generate more sort of interest in it. And so, again, focus on what is it that you're asked to be doing, you know, to do? What are the skills that are needed? But, you know, if you have a LinkedIn alert up now, like, put both on it. I think that's prudent.
Yeah, I agree. It depends entirely on the company. My experience is, I'll call it analyst. You know, most, a lot of companies I see have analyst positions as part of a function. So an analyst within the marketing team or analysts within the operations team, where they'll be responsible for, you know, using data to help improve the operations or just, you know, work within that function. And oftentimes they will have, you know, SQL skills or maybe some other skills like R or Python. But generally, they typically work with, you know, Excel, which is a great, I love Excel. I use it all the time. So that's what I see in terms of analysts. And then separately, I see, I'll call it BI developers and data scientists. You know, there's a lot of overlapping skills there, writing a lot of SQL, of course, and automating the data flow. The distinction being the BI developers are primarily
about building and publishing reports, reports in whatever fashion, whether it's, you know, files or dashboards or what have you. Whereas data scientists tend to also have in addition to that, you know, the statistical modeling and insight responsibilities. So those are the three main areas that I see.
And so it's great. I was kind of hoping that all four of you would land on a very clear, concise definition that you all agreed on and we could solve this problem for the rest of the industry, but it's okay. It's okay. So I hope that everybody at home is like, I think this is all really valuable information. And I think, you know, this advice too, if you've got that that job notification to put both titles in, because it will vary by industry and it's going to vary by company.
Shortlisting candidates
So with that, I would love to kind of think a little bit about how you go about shortlisting a candidate's resume or CV. So in other words, what makes someone stand out to go from that initial, you know, we've got 500 resumes, what do you use to help make that cut or make that decision for who you're going to kind of at least do that initial phone call with?
Yeah, I think that's a really important thing. I think it's different everywhere, right? So some places the hiring manager may have to filter all of the resumes as they come in, whereas in other places, I think there is kind of an HR department or a review team that makes those decisions. So I'd love to hear what your organization uses. And then once you get the resumes, kind of how do you then start winnowing things down?
So I will say, like, we generally will start with a list of skills that we're interested in, depending on the position. And then we look to see if those are somewhat contained in the resume. And it really depends on the type of job. Like, I feel like internships are probably a lot more flexible, right? Where, you know, we have, like, you can go out to our internship website externally right now, and you can see the different buckets that you can apply for for an internship. And so, you know, once you've applied for that bucket, that bucket will go to a review team. And we're looking for some specific different list of skills. And those could be a wide variety of skills, because we have a whole bunch of different managers, right? And so I would say there's differently, depending on the type of position and the type of interview style going on, there's kind of a different process behind each one of those. So that one kind of makes it hard for to enter. I hope that helps.
I'm lucky to have a talent acquisition function that helps me first filter on the technical skills. You know, I tell them, okay, SQL and R or Python or one, you know, or some other programming experience, some, preferably quantitative background, but it doesn't have to be, it could be, you know, a liberal arts degree, I'm a liberal arts major. And then once I see their resume, you know, it's, it's, I don't want to see a four page resume with 1000 programming skills, you know, less is more. I'm looking for depth, you know, have you done a few things really well, as opposed to many things over time, because I'm trying to build a team where, you know, we have a team of experts with different expertise. And so having someone who's really good at one thing is really valuable. So that will complement someone who's really good at something else. That's how I prefer to hire. But obviously, that's just my personal preference.
less is more. I'm looking for depth, you know, have you done a few things really well, as opposed to many things over time, because I'm trying to build a team where, you know, we have a team of experts with different expertise.
I'd say one thing I'd add that I saw on a few resumes I thought was super cool were, I know from an applicant perspective, it can be hard to, to list the programming languages you're familiar without feeling a little fraudulent. Like for me, I'm definitely stronger in R, but I know some Python. But if I just put Python as a language on my resume, it could feel misleading. I'm not like an expert Python programmer, I'm more like basic intermediate. So rather than spelling out all those with the levels, I've sought to just put Python as a level. I've seen like on a sidebar, like a nice graph, that's almost like a bar chart of competency. And so you'd have your top, like that's the full bar, and then everything else would be just arbitrary percentages for competency. It might even be like a markdown template, actually, I think for the resume that it was generated on. So check that out. I thought that was just a nice way to showcase, hey, I'm competent in all these languages to varying degrees without having like a super long detailed section. And it's a nice little graph. So that was cool.
Do you have to specify the scale?
This was not, this is just a visual. So I guess it depends on how, how tight you want to be. That's how you trick people. I'm a D in my R ability, and then everything else is lower.
Yeah, everything else. But if you don't label it, they don't know.
Yeah, I would say like, it definitely depends on a personal perception.
It was a bad joke. More of a rank order relative strength. So it's up to, I guess, you as the interviewer to determine like, okay, what's their 100% at?
With that being said, I'd like to point out like resume formatting, right? Some, you know, some systems will not take bar charts into the resume format, right? And a lot of times they will get printed out on grayscale. So don't make it all nice and colorful. Make, make sure like, if you do use color, that it's going to look nice and black and white. That's just some of my pointers I give to students because some of them get really creative and like, yes, that's, that's wonderful. I applaud your creativity. Just make sure it looks good in black and white.
Yeah, the resume is so important. Just reminder that we have typically three to 500 resumes to get through. So if you submit a novel, it's going to just, it's going to go right past it.
Yeah, that's some other things I say, make sure like you bold stuff that somebody can quickly glance at it in 30 seconds for the first review, right? And you've got their attention, then they'll go back and look at it a little bit more detail if they see something like that really pops out. So I always say like, bold those important things, you know, and make it about you. Because I've, I've had like, students will put an order on their resume. For example, the last class I noticed did a lot of this, like they'll put the company name first, and not the position name. No, talk about you. What did you do? You know, the company is relative, right? I mean, great. This was at Microsoft, but it's far more important for me to see what you did at Microsoft. So I mean, just me personally, like I like that order of like, I can see what you did, and then I'll look and see where you did it at. That's just my own personal preference. But it's like, because I want to know about you. I don't want to know about Microsoft, for example. I'm using that as an example, maybe I shouldn't.
It doesn't have to be a data science job that you've had, you know, if you had some side hustle in college, I think that's really valuable. It shows initiative, for example.
Yeah, we've got tons, tons of great advice. And I feel like we, I'll talk to Chris, I feel like we could do a whole panel on resumes and how to build your data science resume. We could definitely keep this conversation going for another couple of hours. But we are coming up on the top of the hour. And I have been given the responsibility to make sure that we finish on time. So with that being said, I do want to give each of our panelists a minute or two to summarize their best advice for a data science job seeker, or data science manager, and if applicable, to tell us about any openings they may have on their team. And so to avoid the, oh gosh, who goes next? We'll go Jairus, Katie, Iwi, and we'll close out with Rhonda.
Panelist advice and open roles
Yeah, sorry, Jessie, that can you repeat the first question? It was what is the best advice?
What is the best advice? Yeah, so you've got about a minute to tell our amazing community and audience, what is your advice for a data science job seeker or data science manager?
Okay. Yeah, the best thing I guess I would want to see from job seekers that a lot of folks don't, I see not doing, is putting in a little effort to understand the business that they're applying to, how it works, and what their products are. And it's amazing what 15 minutes, either downloading it if it's free, or looking at recent news, or understanding how the business is set up, can do to set you apart from other folks. Because a lot of folks do come in sort of expecting, tell me the metrics you care about, and I'll tell you about fancy stuff you can do with it. But if I am looking to grow you within the role and have you sort of act more autonomously over time, I'd love to see your ability to bring that to the table instead of me having to sort of feed you everything. So there is always 10, 20% of candidates who do that, and I think they get the disproportionate number of job offers out there because of their ability to do so. And again, I don't think it's a ton of effort. So I think that's something that can set you apart. And again, it's a minority of folks who I think do this and do this well. We're filling a job today. I'm sorry, everyone. But I know you mentioned we can do a plug. Yeah, we're extending an offer today. But if folks see positions at SiriusXM, Pandora, Stitcher, or One Big Happy Family, yeah, feel free to reach out. Let me know, and I can give you some advice on what you should apply to, and happy to help you out as I can.
there is always 10, 20% of candidates who do that, and I think they get the disproportionate number of job offers out there because of their ability to do so.
Cool. I actually really agree with that advice. Always come with a list of questions, but I'll try to make this additive. So before I forget, we do have an opening for a senior data scientist. So head out to beam.dental and check out our career site. Would love to hear from you. I'd say probably less so, I guess, relevant at the company I'm currently at. We're kind of a mid-growth stage startup. But for sure, hearing it and some of Rhonda's answers, et cetera, at bigger companies, I think it's tough when you don't know anybody who works at the company, just making sure somebody sees your resume. So I'd say the additive advice I would have maybe is I mean, I am not above, if I'm job searching, stalking the recruiting team at that company. And messaging them on LinkedIn, right? Like just making sure some human sees your resume, I think can be super important if you don't know anybody there, because it would be a shame if it just got tossed out for some silly reason. So I try to be super hands-on with all the resumes we see. So I don't have that great PhD student slip through the cracks, but I think that can happen sometimes just because of the sheer volume of resumes that these teams get and how many openings they have. So try to find a physical human to message if you can.
A quick plug. We have two openings, one for either principal or lead slash senior data scientist position. And there's also one opening for senior BI developer, Amwell, as a telemedicine company. So we have lots of data. In terms of my advice, I think it's important to understand the distinction between having the technical skills to do the job of a data scientist. But then you also need to make sure, and this has been mentioned before, that you have the statistics knowledge as well. You can perform the tasks required, write SQL, write R. But if you really don't understand what it is the models represent, that can limit your growth. So that's my advice.
All right. So I always have kinds of advice. But if you're a female, you can do it. You can apply for data science jobs. Interesting. And my second is always pick something that you're passionate about. You'll be far more interested in it. Don't just pick a job just because it's a data science job. Pick something that you're truly enthusiastic about. And as far as Boeing hiring, we are always hiring. We are a global company with wide, vast, diverse positions open. And you could work anywhere from Spain, Australia, Singapore, to the US, about anywhere in the US you could imagine. Not everywhere, but good, diverse positions. So I'd really encourage you to look at Boeing careers. I forget the exact website, but I always pass it around. Or you can put it into Google and you'll find it pop up. And I would just say if you are looking for a job there, make sure you use a vast variety of different terms. Because for us, I am a data scientist at Boeing. That is technically my skill code. But we also have applied statisticians, applied mathematicians. We have advanced technologists. All of those are similar types of jobs. We even have data science engineering skill code. People who focus on AI, machine learning, autonomy systems. We have library services. We have health services. We have our own fire department. So there's just a wide variety of job opportunities for you at Boeing company.
Awesome. That is amazing. So this is such incredible advice from all of you. Thank you so much for sharing your time and your expertise. So as we close out our session, I of course want to thank you personally. I am going to thank you on behalf of our studio and especially on behalf of our community, everyone who is watching live right now and everyone who is going to tune in over the coming weeks and months. You have been phenomenal guests. And I know I've learned a tremendous amount. And I know that everyone watching has also learned an incredible amount. So with that, our panel conversation has come to an end. And I want to wish you all a phenomenal day. And I hope we can talk again soon. Thank you.