Data Science Leaders | Episode 24 | 25:16 | October 19, 2021
Why It Pays to Stand Out from the Crowd in Data Science
Get new episodes in your inbox
powered by Sounder
Talent is pouring into data science, even though it always seems like there’s not enough to meet demand. Learning opportunities for people getting into the field have exploded in just the past decade.
That means standing out from the crowd—both as a leader and as a practitioner—has become more important than ever before.
In this episode, Bob Bress, Head of Data Science at FreeWheel, explains how professionals at all levels can position themselves to win in a burgeoning market. Plus, he offers advice on how data science leaders can stimulate collaboration and intellectual curiosity within their organizations.
- How to stand out from your peers
- Intellectual curiosity, innovation, and collaboration in large organizations
- Being the CEO of the data science project you’re working on
Hello and welcome to the Data Science Leaders podcast. I am Dave Cole, your host. Today's guest is Bob Bress. Bob is the Head of Data Science at FreeWheel, which is a division of Comcast. Bob also received his undergraduate and PhD degrees from RPI, Rensselaer Polytechnic Institute. Bob, welcome!
Thanks, Dave. It's great to be with you today.
Awesome. Thanks for joining the Data Science Leaders podcast. So on our agenda today, we're going to be talking a little bit about standing out from your peers. So as a data scientist, how do you sort of separate yourself from your peers, both as a leader, but also as a practitioner, as a data scientist? What are some things that you should sort of think about?
And then the second topic we have is bringing data science to the masses. So how in your company do you, what are some tips and tricks and what are some ways in which you can better help the folks on the, let's say the business side or other areas of the company better understand what you do so that you can hopefully work better together, I assume, and collaborate in a better way?
So let's start with the first. So what were pearls of wisdom that you gleaned from your career, Bob, in terms of standing out from your peers?
Yeah, that's a great question. And this is an even more important question today than it was, I would say 10 or 15 years ago, mainly because so many people have gotten into the data science space. If you look out even the offerings from colleges and bootcamps and educational opportunities online, there's so much more available today than there was 10 plus years ago even. And as a result, a lot of people are getting into the space. A lot of people are taking the opportunity to learn data science at one level or another, and separating yourself from the pack becomes even more important than it ever was whereas before if you simply had a degree in it, you separated yourself.
I think there's a couple things people can do. So for folks that are just getting into the space, as somebody who sees a lot of data science resumes, a lot of people tend to have a lot of the same key words, Python, machine learning.
AI. And what you start to realize is all of that is on everybody's resume one way or another. And so as somebody who’s reviewing that, the challenge is how do you now separate the different people, the really skilled ones versus the ones that just have the keywords on the resume?
And so I think there's a couple things we look for. Number one is, can you find a passion in the person for the area that they're looking at? Are they just doing it to get a job and check the box on some of the key skills? Or is it something they're clearly interested in? And there's a couple of things you can look for to see that. Certainly participation in events like Kaggle or blogging about it, or taking on challenges that are above and beyond maybe what the call of duty as part of your job or even coursework would be. I think that shows another level of interest and passion that, an intellectual curiosity really, right? Because in the end, the person you want on your data science team, you want them to be a great problem solver, somebody could throw something at, and they're going to find a way to attack it leveraging their data science skills, but really show a passion for that. So I think finding ways to show your passion for that area and show what you've done that illustrates that, I think helps a lot.
I think also in industry, for sure, you want to see folks that really dig into the context of the problem. So for FreeWheel, we're in the advertising space, ad tech. But whatever space you're in, you want to see that that person wasn't simply applying algorithms and formulas that they've learned in the school and textbooks, but actually digging in, understanding the context of the different problems and taking creative approaches to solving them. Because a lot of times in data science, you're going to be faced with, no two problems are exactly alike. We may be able to take a similar approach, but in the end, the more challenging problems are going to require some level of creativity. So anything you can do to show where you've been innovative and creative and problem solving, I think is a big help.
So a few things to unpack there. Certainly, we talk a lot on the Data Science Leaders podcast about bridging the gap between data science and the business outcomes and making sure that the work that you're doing has measurable outcomes and has high ROI and all that good stuff. So it certainly makes a lot of sense that your expectation of the people on your team is that they understand business acumen and they have an ability to translate some of the work that they've done and be able to describe it in terms of the value prop and the ROI. And along those lines, if they can also show passion for how they creatively attacked that problem, that is a great way to sort of separate yourself from the pack.
But one thing I think is interesting, too, is just that intellectual curiosity, and the fact that if you're looking at a data scientist resume, and they mentioned that they actively participate in monthly Kaggle contests, or even that they've done a few, or maybe they went above and beyond and took a Coursera course or something like that in a particular area of data science to help sort of expand their skill sets. Obviously that shows passion. I'm curious though, in your team, do you do things to foster some of that? So even after they're hired and after they're on the team, if a team member of yours were to say like, "Hey, Bob, I'd love to take a Friday afternoon off and work on this Kaggle problem as a way of learning this or take a Coursera course." Like what is your approach as a data science leader to supporting that?
Yeah, we're definitely a big proponent of training, and that training could take place in a number of ways. Certainly developing new technical skills through Coursera. One of the great things about Comcast and FreeWheel, as part of Comcast, is they offer a lot of different types of in-house training, technical and otherwise. And so there's a ton of opportunity to really take that on. And we've had a number of people do that. And that certainly is good.
I think another area we look to do it and to really challenge folks, but even push the business forward, is through innovation. And we work very closely with an intellectual property team at Comcast to really take the data science projects we're working on and identify what of what we've done here may be patentable? What of what we've done here could be developed as intellectual property that we could benefit from? And what's great about that is I think it really engages folks who are excited about innovation and taking on new concepts. And it sort of challenges them to take a project they've done maybe to another level or to explore some piece of that that we've determined could be developed out as intellectual property. And I think for somebody in the technical space who's really interested in just working on really cool, interesting problems that no one's done before, I think that could be a big motivator. And it is for me and I think it is for a number of the folks in the team. And then when folks are actually filing patents and getting patents granted and getting rewards for that kind of thing, I think it really encourages folks to do more of that.
How many patents has your team filed for? I'm curious.
Oh, over the years, probably dozens. We try to get at least a couple in per year over the team. And we've been lucky, we've had a number granted and we have a number that are in process now, and it's a really great part of the job, actually.
Yeah, most data science leaders probably can point to and say that, oh yeah, they're being innovative. But being able to say that you've patented something really says that you're one of the first to come up with it. It truly is innovative, to be able to say there's actually data to say how innovative we are because we filed four patents last year or something like that.
Yeah, and it's great that people get recognition for that, because it's really a reward on the extra effort they put in to develop out these concepts. On top of that, of course we want there to be a great business application for it, but that comes with the territory.
Yeah, well that's great. In terms of training, you mentioned Comcast, the parent company of FreeWheel, has a lot of in-house training. Do you leverage external training as well? Do you encourage your team members to take Coursera type courses and things like that?
Yeah, definitely, definitely do that, participation in conferences and things of that nature, keeping up to date on the latest developments. It can really take form in a number of different ways, depending on how that individual wants to develop, but something we strongly encourage. And I think today, in particular, as cloud computing and software applications for AI and other data science are constantly changing and coming out, folks want to stay up to date with their skills there. And anything we can do to encourage that, we want to, and certainly we're the beneficiary of that because inevitably we'll find areas where we can introduce that into the business to solve new problems.
Yeah, it's a win-win. Certainly it's hard in most roles. And obviously I'm a bit biased here, but in the data science world, things have just, if you look at how data science is done today and you compare it to how it's done 10 years ago, you look at the advent of cloud computing and how that's impacted in a positive way, the ability for data scientists to be able to do their work, just as an example. This open source, there's data science platforms, there's all sorts of, Spark, there's all sorts of different tools and technologies that have helped data scientists be more productive and in their models, even better and more accurate and things like that. The world is constantly changing, and I think data scientist is in the top decile, at least, of roles, of jobs that really require constant training. And if you have that mindset as a data science leader and you really espouse it, I think it just makes your team that much better. And I'm curious, I'm going to put you on the spot here, Bob. But have you gone back to training personally? Have you done some of this, too? Are you hands-on?
Yeah, absolutely. I think the reality is, we want to have a very technically skilled team in the data science space. But to manage a team like that, you have to be able to have intelligent discussions around the tools that they're leveraging, the methods that we're using and what the alternatives are. And I, myself, a lot of times I'll do it through a book around machine learning, AI development, just to try to keep up to date with different tools. And I try to allot at least some time per week just for that, sort of sharpen the sword, you might call it, just keep the skills up to date. And it's harder now than I think it's been in past years, just because things are moving so quickly and there's so many different options and areas where you can learn and really dig in. But, yeah, super important to do that.
Great, and one last thing in terms of keeping your skills up to date and sort of helping yourself and not just you as a data scientist sort of separated from the pack, but also your team itself separate themselves in terms of just you being a high caliber, high performing team. We talk a lot on the Data Science Leaders podcast about collaboration. I imagine that there's formal ways to brush up your skills, but also there's the informal ways of just collaborating and working together. Is there anything unique that you've done with the team to foster that collaboration?
Yeah, one of the things we try to do often, and it's been more challenging under a pandemic, but Comcast is a very large organization and FreeWheel is a part of that. But in the broader organization, there's a number of data science teams. There's a number of teams that are doing really interesting work in that space.
One of the things we've tried to do is connect those teams as much as possible, even if it's just informational. So before the pandemic, one of the things we did at FreeWheel is we held a data science day. We invited different data science related teams from across Comcast and NBCU to come and present some of the coolest projects that they're working on, really with the intention of just connecting, connecting people. If nothing else, it brings an awareness of new approaches and problems people are tackling. But it also helps you make connections so that when you're doing something fairly similar, you have somebody within that same umbrella to reach out to. So a big part of that is just connecting, sharing information, best practices.
Even more recently, we had in a team meeting, there was a data scientist at Comcast who had written a book, and we had him come into the team meeting and tell us about his experience in writing that book. So we try to stay connected as much as possible and looking forward to sort of getting back in person so we could do that.
Yeah, so in this data science for the masses day, what was it the formal agenda? Did you have folks who presented topics as a way of sharing? Or is it more informal get together happy hour type?
Well, I think a little bit of each of those. So there was a happy hour to end the day, but we did reach out to different teams in advance to say, "Hey, is there a project you're working on you would love to sort of present to this broader group? I know there's a number of people here who would love to hear it." And that's what they did. They would sort of preschedule. Everyone would get 20 minutes or so. They would highlight some of the best applications that they're working on. It was just a great way to hear and see some of the interesting things that are happening.
I'm of the opinion, Bob—and I won't apply this to your parent company, I'm sure this doesn't happen there—but I'm of the opinion that at large companies especially, that there are pockets of data science teams that are working on very similar if not identical problems and challenges, and there's a lot of duplicative work. I'm not going to put a number to it, but I'm just of the opinion that there's a lot of that going on and not enough collaboration is going on, not enough partnership. And so much of data science is about brainstorming, is about spitballing, and experimenting. And it's hard to do that by yourself as a lone wolf out there creating your own models. But if you have the partner out there maybe in another division, but has very similar background and has an expertise in deep learning or in forecasting, or marketing use cases, or what have you, and you're able to foster that relationship beyond just a once a year event where you get together...I know that when I was doing work myself, that was always helpful, having somebody to bounce ideas off of. So I think that is just so, so critical.
Yeah. No, absolutely. And different teams and organizations are at different places on the, I'll call it a data science maturity curve. And over the years have sort of refined capabilities, whether it's how they process data, clean data, set up workflow, automated workflows. And for teams that are on the sort of earlier end of that curve, they can really benefit, they can almost skip up the curve by engaging more with some of the teams that are further down that. And so that's where a lot of that connection can really be beneficial.
Yeah, I wonder if you've ever entertained, I'm just going to throw this out there, the idea of having people join teams for a short period of time. So having folks just work on your team for a month or two, just to understand a little bit of your process and sort of work on loan, go back to the other team. And I don't know, have you ever tried something like that?
Yeah, we've had folks, there's folks even around the company that are not necessarily in a data science role, but are interested either exploring it or developing themselves in that space. And we've done what we'll call a gig where somebody can dedicate, it could be a number of hours, a week, where they'll spend some time on a problem we're working on. So it'll help us out because we have the extra hands, but it also helps them out because they're developing skills under a mentor who can guide them. So we have done things like that. I would say that's something that's fairly common is, well there are so many people who are interested in getting into this space that I do get pinged a lot about, "Hey, what are some of the best ways to get into this? What are some things I could read or study or online courses I could take?" And we see that a lot.
Yeah, I'm sure. Well, this has been, a lot of good insights there in terms of ways to collaborate within large organizations, ways in which to sort of separate yourself from the pack through training and learning. And also, when you're putting that resume together and applying for a job, make sure you highlight your passion and your curiosity. I think that's what's really important from a data science perspective. It's not good enough to just be able to talk the talk and talk shop and talk algorithms and talk approaches, but also being able to truly understand the business and also talk about what's current in the industry. What are some of the trends that you're seeing and some things that you might want to experiment with if you got the job, what approaches you might want to..." I'm curious, go ahead.
I was just going to say, one of the questions I love to ask is, "What interests you the most? If you had a job in data science, what are sort of the ideal things you would be working on?" What would really drive you to be interested in that?" And that's always great to hear because if a person is really aligned to what they love to be doing, that's when you really get some great results.
Yeah, a hundred percent. Well, hey, the one question I tend to ask from time to time is if you were in a time machine, Bob, if time machines existed, and you were able to go back in time, say 10, 15 years early in your career, what advice would you give yourself?
Yeah, that's a good question. Sort of along the same lines, I think really digging into not just the data science aspect, but the business context. And I think part of that is really see yourself as an owner of the problem that you're faced with. You're the CEO of this data science problem that you're faced with. Really take ownership of it. And when I say that, it's not only sort of see your piece of it through to success, but also start to think about for this project to be a success, what else has to happen? Who has to be engaged? Who has to believe in the work that we're doing and what can we do to help convince them to do that? So I think one piece of advice is just really be an owner of the project that you're on. Be the CEO of the project you're on and really extend yourself beyond the technical aspects. So that's one thing.
I think another is really around connecting. So in data science, as you come into data science, you want to work on really interesting and cool problems. And if you're really lucky, those will come to you because somebody who knows enough that data science can solve this problem can bring that forward to you. But that's not always the case. Some of the most interesting problems come when a data scientist says, "Hey, I recognize this as something that can be solved with the tools I have in my toolbox, but I know enough about the business context to now apply that when nobody was thinking of that before." Whether it's a unique way to use classification or forecasting or some unique combination of data science methods that nobody's applied before.
So I think the way to get exposed to that is to connect with as many folks across the business as you can. And even if it's just understanding, what are the problems you face? And some of those problems that different groups face are going to be something that data science can really make an impact. Not everything, but when you're in a position where you can identify those, bring those forward and solve them when, when that hadn't even been done before, that's when you can make a real impact.
And then on the technical end, I think it is really around constant learning. So I think staying up to date, trying new things and not being afraid to fail with new and different approaches. Take some risks and then be vocal about it. And not everything you do is going to be a home run, but I think when you put yourself out there and really extend yourself and try new things enough, you're going to see some success.
Yeah. I think there's a lot of good advice there. I especially like the idea that I think oftentimes data science teams see themselves as folks are waiting for the business to come to them with their problems. And a high performing data science team is not just waiting for those problems to come to them, but also proposing solutions to problems that they are aware of. And that requires them to really understand the business aspects. If they're seeing, like I said, some new approach and they know of a problem that it can potentially solve on the business side and you, yourself, and team member go and present to the business, you're going to make fast friends with the folks on the business side, for sure. They will see you as a partner and not as order taker of problems of theirs. So I think that's sage advice.
And also, staying current. And we talked a lot about that already in terms of training. And I think all data sales out there should be hearing the types of things that you're doing and doing something very similar in their role. If anything, to not just stay current and stay sort of cutting edge, but also to keep your team interested. It's a great way to retain employees. I've always believed that if you're not learning, you're leaving. I think you can pay people gobs of money to stick around, but eventually they'll leave if they're bored. It's hard to pay people to sit around and be bored. So keeping them educated and keeping them challenged is always a great way to have your high retention with your team.
So, Bob, this has been an absolute pleasure. I really appreciate you coming on the Data Science Leaders podcast. If people want to reach out to you, can they find you on LinkedIn?
They can find me on LinkedIn. You could just search for Bob Bress. You'll see me there, and happy to connect with any of the listeners and hear about folks’ experience.
Fantastic. Well thanks, Bob, and have a great rest of your week.
Thanks for having me.
Listen how you wantUse another app? Just search for Data Science Leaders to subscribe.
About the show
Data Science Leaders is a podcast for data science teams that are pushing the limits of what machine learning models can do at the world’s most impactful companies.
In each episode, host Dave Cole interviews a leader in data science. We’ll discuss how to build and enable data science teams, create scalable processes, collaborate cross-functionality, communicate with business stakeholders, and more.
Our conversations will be full of real stories, breakthrough strategies, and critical insights—all data points to build your own model for enterprise data science success.