Economics & Sociology at IBM (feat. Ben Zweig & Jonathan DeBusk)
Alumni Aloud Episode 15
Ben Zweig and Jonathan DeBusk are alumni of the Graduate Center’s PhD Programs in Economics and Sociology, respectively. Both now work in IBM’s Chief Analytics Office. Ben is a managing data scientist, and Jonathan is a senior managing strategy consultant.
In this episode of Alumni Aloud, Ben and Jonathan explain what their jobs look like during a typical work week. They also give practical advice to those who are considering using their data skills in a corporate environment.
Listen to the episode below, download it, or stream it in Apple Podcasts (or your preferred podcast player).
VOICE–OVER: This is Alumni Aloud, a podcast by Graduate Center students for Graduate Center students. In each episode we talk with a GC graduate about their career path, the ins and outs of their current position, and the career advice they have for students. This series is sponsored by the Graduate Center’s Office of Career Planning & Professional Development.
ABBY TURNER, HOST: I’m Abby Turner, a PhD candidate in Educational Psychology at The Graduate Center and I work here at the Office of Career Planning & Professional Development. I interviewed Jonathan and Ben after they finished conducting interviews with GC students in our office. It was a perfect time to get their advice on interviewing and working in business after graduate school. I have Ben and Jonathan from IBM in the office and they’re going to talk to us all about their careers. And they came from two different programs so we’re going to hear about that. So Ben why don’t you tell us where you came from in The Graduate Center and what you do at IBM.
BEN ZWEIG, GUEST: Yeah, sure thanks. So I came from the Economics department, I graduated just about 3 years ago and now I am a Managing Data Scientist in IBM’s Chief Analytics Office.
JONATHAN DEBUSK, GUEST: Okay I am Jonathan DeBusk and I graduated about 5 years ago from Sociology at The Graduate Center and have been with IBM about 5 years. And I am a Senior Managing Strategy Consultant within IBM. And one thing to add to that, I guess we’ll go into it a bit more, but Ben and I are internal consultants at IBM, so our clients are IBM executives.
TURNER: Great, so since we’re on that topic why don’t you tell us what does it look like working and doing the job you do at IBM. What does the day-to-day look like? Either of you could take that.
ZWEIG: Yeah sure I’ll get started. So as Jonathan mentioned we’re an internal consulting group, so basically our clients are internal IBM executives that have lots of data and lots of problems and they don’t know what to do with it. So we basically analyze their data. So you know we get access to databases, we pull data into our computer, we open some programs like you know Python or R or something, and then we produce some descriptive statistics and charts, etcetera. And we come back to the client and they tell us you know what they need, the problem they need to solve and then we build some models for them. So we predict things or we develop some set of recommendations based on analytical models and we communicate these to the client. And in doing so it’s kind of an iterative process where they end up telling us what they need and we end up telling them what we think they need. And at the end of the day, we build a pretty robust model, pretty robust set of recommendations to help improve the IBM business.
TURNER: Ok Jonathan anything you want to add?
DEBUSK: I think Ben said everything. The only thing to just emphasize a bit more is our job is not just data scientists or researchers. I think Ben said this but you know half of our job is being a consultant, where we go out and we help the client identify the problem and help them translate the analytics into something actionable. So part of our job is finding those recommendations but also making sure they’re implemented. So it’s not just a researcher doing research, but making sure the output gets implemented.
ZWEIG: And that’s actually the hardest part.
TURNER: Ok. So does the day-to-day change then for you or do you have a standard kind of schedule and routine?
ZWEIG: I would say the day-to-day is pretty different but the week-to-week is more consistent.
DEBUSK: That’s what I was going to say. *laughs*
TURNER: Ok, could you explain that a little bit more?
ZWEIG: Yes so I would say we have a fair amount of meetings right. What percentage d you think is meetings?
DEBUSK: So there are all types of meetings. I would say you know it week-to-week it could be anywhere from like 10 to like 40 percent depending on the time. But most of our meetings are kind of in this consulting perspective where it’s when we’re meeting with our client to kind of present what we’ve done or understand their problem. And then we have a lot of informal meetings with our group, our internal group, to kind of understand how we’re going to answer their problem, where are we going to get the data, work with each other to understand the model that we’ve just built. So those are two main types of meetings. We also have meetings within our group, you know larger all-hands meetings for development sessions. But I would say kind of an average day-to-day is getting together with our team to figure out what we’re going to do that day, that week, figure it out and work on it. Then maybe a meeting an hour later to figure out how to get some data from someone then informally meet to go over a model, meet with a client and then kind of work independently or with the group throughout the day.
ZWEIG: Yeah I mean there are some days where that just get kind of consumed by meetings. Things pop up and you know I’ll end up getting very little work done. But then there are other days where there aren’t so many meetings and I can actually do you know kind of heads down work for a good 5, 6 hours, that’s probably the max per day. Yeah I would say it’s a good mix of meetings, actual work and kind of brainstorming.
TURNER: Okay so you’ve already mentioned R and Python. So do you recommend these two programs as the ones that IBM are most interested in you knowing?
ZWEIG: It’s not specific to IBM. But I would recommend these tools for someone who wants to get into data science or analytics, especially Python. I mean it’s the most popular, most well-supported program these days. I would definitely recommend not spending time on programs like SAS, Stata, I would say SPSS but it’s an IBM program. But yeah I mean the open-source programs are just getting so much attention, so much traction, they’re getting so much better than their competitors. And it’s really just, you have to know them if you want to do any sort of data analysis out in the world.
TURNER: So where did you guys learn R and Python? Did you learn at The Graduate Center?
ZWEIG: The internet.
TURNER: Okay tell us about your learning process through to picking it up, because we do have a lot of graduates telling our current students that this is an important skill if you want to get into it.
ZWEIG: It’s so critical.
DEBUSK: And I think the pace at which it has changed is also very critical to understand because it really is just in the last year or two where that’s become the main skill set. I think whenever we hire or interview, that is basically one of the main things that we look for is the Python skill set. When I was applying for IBM, it was 5 years ago and when I came in knowing SPSS from The Graduate Center was what helped me get the job. And I think it’s been said that those kind of programs, although may help one, I think you really need to show some kind of open-source coding ability.
TURNER: Okay, so how did you guys pick up your coding ability?
ZWEIG: Yeah so I had a similar a similar experience as Jonathan in that I did’t know these things going in and that was 3 years ago and I knew Stata, it’s the most popular too in economics. But you know it’s not used in industry, it’s expensive, it’s not open-source, it doesn’t have all the capabilities and packages etcetera. So I learned Python from Coursera courses. There’s you know a million great resources out there, that’s pretty easy.
TURNER: So did you learn it after you were hired or just do it before?
ZWEIG: Yes I learned these programs after I was hired.
TURNER: Okay but you did get hired knowing Stata?
TURNER: Okay so that probably helped as well. Okay Jonathan, how did you learn?
DEBUSK: So I think the main way that people in our department have learned is through Coursera as well as IBM in our group in general is very good at kind of educating and developing those within our group. So we have development sessions where the whole team is kind of as a group learning and moving with Python. But I think, as Ben said, one very easy way for someone to pick up is on something like Coursera.
TURNER: Okay so outside of the programming skills, what skills did you pick up at The Graduate Center during your time in graduate school that are really transferring well into your career now?
DEBUSK: So I think we all kind of know for a job like this it’s very important to have deep statistical, analytical skills and now Python is kind of the main way to go about that. But I would say the one that I think is very important that people may not really selling very well that can sell very well is teaching. Because as graduate students we all teach and that gives very valuable experience in terms of you know being able to communicate in front of a group but also being able to take any kind of question that could be very complicated, ordinary not complicated, and be able to answer to that group and keep a group engaged. And I think for us coming into this environment where you know our job role is as a consultant. Being able to work with clients, to take their problem and translate it into something actionable, I think that is the skill set that really helped. So working with clients in meetings, interacting with people who do not have such a technical background because you forget that people don’t. So you know we can communicate very technically which is good but the more important skillset I would argue is being able to translate those technical skills into something that a non-technical person would understand. And again I think teaching really helps do that.
TURNER: And how do you translate that onto your resume about your ability to communicate to non-technical audiences?
DEBUSK: Well I think actually you raise a good point. I haven’t thought about you know how you exactly spell that out on a resume or CV. But I think having it on your CV as work experience, the teaching and that kind of listing the classes is something one should definitely do. But I think it more so comes out of the interview process that making sure as a graduate student you sell that. That you say, “I have this you know maybe this research position but I also have teaching experience which…” And then go into how that prepares you.
TURNER: Great. Anything to add, Ben?
ZWEIG: Yeah sure. So I really agree with Jonathan that in order to be an effective data scientist you really have to be an effective communicator. And teaching is an experience which helps you distill complicated ideas into their essence and that’s really a lot of what we do day-to-day in order to actually get our models implemented. Another skill that I think I want to touch upon is, you know as social scientists we think a lot about causality and policy. And I think it’s a little bit unfortunate but these skills are not really stressed in the world of data science yet. So I think it’s something that data science could learn from social science. You know we think about how to build models how to construct our data and our algorithms such that we can have an interpretable, causal effect and I think that’s really not used.
I mean I’m kind of repeating myself now. But I think that’s something that in the business world people really want to see. You know it’s not sufficient to get a really good prediction. And I think people that come from let’s say computer science, engineering, math, do tend to think in terms of predictive accuracy. But when you’re communicating with a business leader who’s making policy, they really want to know okay what is the impact of pulling this lever. And that’s something we really have to be willing and able to quantify. So I think techniques that we learn social science like instrumental variables, regression discontinuity, fixed effects, you know these are these are methods that can be deployed in practice, that really help social scientists become the best data scientists.
TURNER: Great okay, interesting. Can you compare the speed at which the business world works to say academia or what it was like in graduate school? Any comments on that?
DEBUSK: Yes so I have a comment on why I decided to go into the business world because of that. And actually a skill that I didn’t get in The Graduate Center that helped me with IBM… So I realized when I was halfway through my doctoral program that I really liked working with data and I had various like research positions and worked at think tanks. I liked taking data and solving problems but academia moved too slowly for me. Either no one was going to look at what I did, no one really cared after it was done, but I liked working on the data. So I started to realize… I went to a panel here on careers outside of academia and met an alum that led me to this job. But through talking with her I realized that in the business world you know people actually care about the results because they take action on it. And you can be judged on that action and you’re compensated according to that.
So it moves a lot faster and people actually are very interested in what you do. But I think a skill that I didn’t get in academia that I picked up or realized I needed to pick up while I was in IBM was this idea of, maybe something doesn’t have to be perfect. Because I think as graduate students, we work on something for months and months to get you know to that .001% you know you could spend a month getting there. Whereas in the business world maybe it’s not exactly right, but you have two days to work on it and there’s a deadline. And we find whenever we interview or bring on people is we have to change that mindset, that it doesn’t have to be perfect but you know let’s get something out there and the analytics can help. I think that’s something I had to learn.
TURNER: Great, that’s really good advice. Did you want to add to that Ben?
ZWEIG: I think I agree with Jonathan. You know the stakes are high in the private sector and you know robustness checks don’t matter, p-values don’t matter.
ZWEIG: Yeah I mean statistical significance is such a weird concept because people that know a lot about statistical significance don’t trust it and people that don’t know a lot about statistical significance don’t care about it. So it’s kind of useless no matter how you look at this. I mean you know as an academic you can’t really convince someone by showing them your p-values you know, it’s kind of laughable at this point. And in academia people want to see an impact that they can act upon, that can get them excited. And it’s really about the economic significance of doing something.
TURNER: Any comments on how you built that skill to meet these deadlines and kind of switch your pace from graduate school. What was that transition like?
DEBUSK: To be honest I think it was being there, being in that environment and just being trained. I mean that’s all I can say is that’s how I saw it and that’s how I picked it up.
ZWEIG: Yeah I think just having a new job in general is exciting you know. You want to move fast, you want to be impressive and so yeah I think it’s pretty easy to adapt.
TURNER: Great. And at IBM when you start with this company is there a certain training you go through? What’s it like for a new hire?
ZWEIG: Yes and no. I mean I think there’s a few days of just getting to know the systems, the acronyms, which there are many of. You know it’s a new language a little bit. But I think for the most part, you’re assigned to a team and you have to start producing something, you have to do something that produces value. And then as you say the more these meetings you get to know what the priorities are and you pretty much learn it as you go. Pretty much hit the ground running.
DEBUSK: And in the Chief Analytics Office, there’s about 70 of us total and we work on teams of like 3 to 10 people. And I think as Ben was saying, it’s a very collaborative group so whenever you get on a team depending on whatever project that is, the team works together and kind of moves as a group. You know if we need to learn this new method, they figure out who’s going to pick it up and teach it to the others. So your first few weeks or months is a very collaborative environment in that sense. That being said there are also within our group formal educational trainings that we all go to once a month or two. And IBM is just in general such a large company, you know it’s the largest private research institute, so they’re very committed to their employees developing. So there are tons of opportunities within IBM for development.
TURNER: So you’ve both most mentioned working in teams and you were saying it usually ranges from 3 to 10 people. How did, or did, graduate school prepare you to be so collaborative. Working on a dissertation is a very independent endeavor.
ZWEIG: I would say grad school did not really prepare for that. I think what’s interesting about that the team that we’re on and I guess data science teams in general is that they’re very interdisciplinary. So you know we’re both from social science but economics and sociology are fairly different. We have people on our team from operations research, math, physics, computer science, etc. And all these people come in with different terms for the same thing so I think it’s really challenging to talk to someone from a different field. You’re both talking about the same thing but you’re using different words and you’re kind of talking past each other. So I think that’s very hard in the beginning, whereas for me if I’m working with other economists we’re like speaking the same language, we’re like really making progress quickly. And it’s a little bit slower in an interdisciplinary environment which can be frustrating, but I think also incredibly rewarding because you know they come in with different perspectives, different ways to approach a problem that are often really valuable. So it pays off.
TURNER: Does IBM look to… it seems like even from The Graduate Center we have graduates from these very programs… do they look to diversify or is it just a coincidence that you guys are all from different programs?
DEBUSK: I think IBM in general does. I think there’s actually kind of a campaign that IBM’s thinking about right now about looking for like data scientists from non-traditional field. Because people generally think of computer science or something like that as being where data scientists come from, and IBM is trying to expand and say it’s not just computer science but also social sciences. Which again Ben and I are very strong on, that we think that those skill sets are exactly what’s needed for data science. But in general idea IBM has always been one who looks for someone from a non-traditional background to fill various roles.
TURNER: Great, great. And is there any room for the humanities at IBM? Can you tell us about that?
DEBUSK: Yeah so, names change at IBM all the time, but there was a group called Client Insights, which was the group that did analytics on our clients to understand how we go after different clients. And the head of that department is a vice president who studied theater. So prime example of a non-traditional field leading in an analytics role.
TURNER: Great, awesome. So what advice would you give to current students who are thinking about going into private sector, business, maybe something involving data? What should they be doing now?
DEBUSK: So I think we covered you know the skills that you need in terms of kind of Python, understanding data, to sell the things like teaching. There’s a good amount of advice I think I probably wanted to give in more so kind of doing the interviewing. So one thing is in the business world I think if you don’t give a one-page CV, you can often get just overlooked or thrown away you know. Particularly if you’re a PhD coming into the business environment. One- because they think you just can’t summarize what you’re doing on to one page, you know you’re a long-winded PhD. It seems like a very small thing but the ability to summarize your work into one page is very important. I think when interviewing, just remember you know if you’re going into a business environment, it’s different than academia you know. So things like being prepared for a business interview just in terms of you know even dress, style, the way you communicate with the interviewer. Making sure that you know they see that you’re someone who can be succinct, that you’re someone who is energetic about what you’re getting into.
TURNER: A certain level of enthusiasm about what you do.
DEBUSK: Exactly and making sure you understand that it’s a different environment than academia. And one thing also to add on that is very important I think for PhD’s going into the business world because we’re not really trained on it are these kinds of business case study questions. So we always tell people you know if you’re hired into this position, you’re going to go through a few interviews where you get these style questions. And to kind of prepare for that you know it’s like McKinsey, Bain or BCG. Just kind of Google McKinsey business case study questions.
TURNER: Ok because I don’t know what that is.
DEBUSK: So it’s a formulaic way that in the business world they interview where they simply give data and they ask, you know why is the revenue in this company going down? And it’s an interview style where it’s all about acting with the interviewer. But anyhow if you’re not familiar with that and you’re going for the business world, I would suggest looking into that. So those are my pieces of advice.
ZWEIG: Yeah I think those are spot on, I just want to second them. So I think in interdisciplinary teams you know PhD’s are often competing against MBA’s. And MBA’s are almost unfairly good at interviewing because they spend so much of their time prepping for it. And interviewing really is a study-able skill. But it takes a long time, I think MBA’s put in like 100 hours of interview prep. And it shows, I mean they’re amazing. They can talk about things, they’re concise, they listen well, which is a skill that I think is very tough but noticeable when you’re interviewing someone, whether they’re actually listening to you or not. So I think doing interview prep, doing these case studies is really helpful. I mean even if you’re going into a field that is not doing case study style interviews, knowing the case structure how to approach interviews in a case study process is worthwhile.
So you know there are lots of books on that, there’s Embrace The Case, there’s A Case In Point, there’s a bunch of things. You can read all these and they’re pretty easy to understand. Another piece of advice that I would stress is networking. I think it’s worthwhile to network with people in the field just to know the language they use and just you know practice talking about yourself easily and hearing about them, learning about their experiences, I think it’s just a good skill. And yeah I mean this issue about their language is really an important one that gets lost. I mean especially in academia every discipline has their own language. And I think the business world has its own language and there are trends, there are things that are the hot new thing in business that year. And knowing those and being able to speak knowledgeably about those I think is really critical.
TURNER: So what’s the best way to practice this interviewing for a PhD student?
ZWEIG: Reading the books and just getting coffee with other professionals.
TURNER: Maybe going to the career center…
DEBUSK: Yeah I know Jenny Furlong has case study questions, has the books on those, sometimes hold events. And I’m sure if you reach out to her she can definitely help with the interviewing process. But I think as Ben said, find a friend who just got a job recently in the financial sector or something and they can definitely help you.
TURNER: Great, great. When you started talking about CVs, I thought about the question of, how important is it for someone coming out of graduate school to have non-academic references if they’re applying to private sector? Is that an issue at all?
DEBUSK: I don’t think it matters.
ZWEIG: Yeah I’ve never looked for a reference on a CV and I don’t think I’ve really noticed it on anyone else’s CV.
TURNER: You guys don’t use references or you don’t handle that work?
ZWEIG: Maybe the HR department may look for references but as Ben said I’ve never even looked for them, we’ve not contacted them.
TURNER: So you don’t need references?
DEBUSK: Particularly from you know MBA-style CV’s, which would be the style that all should have is a one page and I don’t even think they have references on that.
TURNER: You guys keep mentioning a one-page CV. I’ve thought it’s normally resume or CV.
DEBUSK: We use resume and CV interchangeably. We want a one-page resume highlighting education, work and skills I think are the things. And always kind of rule of 3, like if you’ve had 8 work experiences put them into the 3 and then make it really actionable. So you know the work experience- your title, the position, the place and really answer the questions of “what did you do, how did you do it, and what was the impact?” Make every work experience like that. And I think instead of kind of references, make sure that you have work experiences that may be outside of your field. Like if you’ve actually not had any research experience, take whatever work experience you have and at least turn that into something to show that you solved a problem somehow, you had some impact.
ZWEIG: Yeah I would add to that and say tools and skills I think are critical to emphasize. So you know we want to know if you use C++ or Java or Python or whatever. You know these are useful to know because they’ll jump out. If you have 10 seconds to look at someone’s resume, that’s a go to section. One thing that you could definitely skip is papers that you’ve written and posters you’ve presented, conferences you spoken at. They’re kind of meaningless. Because you know especially in disciplines that are not your own discipline, you don’t know how important a conference was. So it’s just a thing that people gloss over.
DEBUSK: Yeah I completely agree with Ben. And I would say at the bottom like have the skills section and it’s really important to list Python or whatever you know. And if you want to, underneath that if you’re very strong about communicating the fact that you have publications, whatever, you could have a line like you know “Publications, conference talks, whatever can be provided upon asking” or whatever the wording is you’d use for that.
TURNER: Can you guys think of anything else that you’d like to share either about getting the job maybe for people thinking about this area. How would they make that decision?
ZWEIG: I would say you know going into data science from academia is a really great transition. I would wholeheartedly recommend it. I think academics play a very important role on data science teams. You know I think there a lot of people who can kind of implement models but I think a really profound understanding of the statistical foundations is important. And I think we learn that in academia. I also think that being able to do research and think in an innovative way and realize you know methods that are flawed is important. And I think tech is very exciting, there are so many opportunities to actually start producing value for companies that are making things. IBM as an example, you know IBM is producing software and services and delivering value to people. And we can actually be impactful on the margin. Well maybe a little more than on the margin, but we can we can actually be impactful. So I think it’s an attractive place to look.
TURNER: Yeah, you were just making me think of…maybe this is a better question. What does your PhD bring to your job and or maybe to the work environment for your colleagues at this point?
DEBUSK: I think within IBM it’s seen as a credential. PhD’s are really respected and looked up to and seen as an SME, subject matter expert. So I think once you get into IBM it’s seen as a credential, it’s respected and it helps you get in. And I would say kind of going back to the transition into that is one… if you’re taking your PhD into the business world, just remember it’s not academia so you have to think about yourself differently and the way you interact with others. So you have to sell yourself a little, you have to be energetic, excited. Reach out to people. For us in particular, when we look for a PhD it’s not one who just kind of sits there at a desk and does something very smart and waits for someone to come and tell them what to do. It’s about being able to communicate and get out there and show people that you’re excited and that you’re not just you know a PhD who just sits in an office, who is as smart as possible but no one really knows because you don’t get out there.
ZWEIG: Yeah, I think to that point there are some traps that PhD’s fall into. I think you can tell when we’re considering hiring someone, you know it wouldn’t be uncommon or surprising for someone to say, “oh yeah but are they a little bit too PhD.” And I think what they mean is you know, are they just trying to do math by themselves or do they not care about the business problem or are they you know a little head in the clouds? Yeah I think it’s important to combat that when you’re going out on the job market. You know don’t talk too much about your research and try to talk about new things that you don’t know so much about. Try to talk about how you could apply your skills to solve different types of problems. Be more action-oriented because PhD’s have a reputation for being not action-oriented.
TURNER: And so your actual field of study even specifically your dissertation, is it going to mean much after you get that degree right?
DEBUSK: Nope. Not really no.
TURNER: Don’t care about specific research.
DEBUSK: I think what Ben just said and kind of a culmination of what we’re saying is that PhD’s learn a lot and become very smart on something and that’s why companies want them. But they want the PhD to be able to understand their value. So the value of a PhD is not sitting in a room and doing something, working on it forever.
TURNER: It’s the skills you built while doing that.
DEBUSK: Yes. Being able to take those skills and make them actionable. So what we talked about earlier, your time box. They want an answer from you in two days and being able to understand that you’re smart enough to do something in two days.
ZWEIG: Yeah and one caveat to that is that there are groups there are research groups where companies actually do hire PhD’s to sit in a room, publish papers and work on a very small problem. And a lot of companies, a lot of tech companies, have dedicated researchers. And that really is an academic position. I think what we’re talking about is a position within the business.
TURNER: Gotcha, okay. So you guys have explained really well how to get into the business world with your PhD. So I want to thank you so much for taking the time to sit for the interview and we’ll talk to you guys soon.
TURNER: Thanks again to Ben and Jonathan for talking with us today about their careers at IBM. If you’re interested in connecting with alumni who got careers outside of academia like they did you should come to one of our upcoming events this semester. Check out our website at cuny.is/career plan or follow us on Twitter @CareerPlanGC for details on our programming. We hope to see you soon.
This entry is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International license.