(Music)

VOICE OVER: You’re listening to Alumni Aloud, a new podcast by Graduate Center students for Graduate Center students. In each episode, we talk with a GC graduate about their career and the advice they would give current students. This series is sponsored by the Office of Career Planning & Professional Development at the Graduate Center.

(Music)

ANDERS WALLACE, HOST: In this episode, we speak with Ivo Vigan, an alumnus from the Graduate Center’s Department of Computer Science about his transition from academia to working as a computer programmer and research analyst for Bloomberg, L.P., a financial software, data, and media company based in New York City.

Ivo talks about running his own digital startup while still a student at the GC, how he made up his mind to leave academia, and how working in the private sector can be surprisingly similar to life as an academic. He also talks about how his PhD gives him an edge and the kinds of skills that create value for Bloomberg. He’s interviewed by myself, Anders Wallace, a student in the anthropology program here at the GC.

IVO VIGAN, GUEST: Yeah, hi. So, let me thank you for inviting me to talk on your podcast. So, yeah, my name is Ivo Vigan. I grew up in Zurich Switzerland, and I’m currently working at Bloomberg as a research scientist in natural language processing.

WALLACE: And your academic background, you’re a graduate from the CUNY Graduate Center in computer science?

VIGAN: Yeah, exactly. So, I did my Masters at ETH Zurich in theoretical computer science. I continued my studies at Grad Center of CUNY in what is called computational geometry. I graduated last May, so May 2015.

WALLACE: So, can you tell me how did you come to do the work you’re currently in?

VIGAN: I guess every student when you get to the end of your PhD, you think about should you stay in academia, and you wanna do a post-doc, or should you go to industry. During my PhD, I went back and forth between stay in academia and going to industry. I think the event which ultimately made up my mind for leaving the academia was during my PhD with two of my colleagues, we had a startup for about half of the year where we worked full-time on it, and I really liked the speed and the energy you create in this startup world. It’s fast-paced, it’s short-term goal oriented.

And sometimes, especially in more mathematical fields, if you submit a paper to a journal, it takes two years. Two of my papers are still not reviewed right now, so it’s a much slower process, and I think ultimately, I really realized I’m inclined way more towards the fast-paced environment. The recruiter contacted my professor and she was looking for so-called algorithms engineers. At that point, my professor really wanted me to do a post-doc, but the good thing was that he then forwarded me that email, and ultimately, basically I got the job through him, so he couldn’t be mad at me for not staying in academia.

WALLACE: What was the job?

VIGAN: So, this was a startup in New Jersey called Driversity. So, their idea was that you can use smartphones, especially smartphone sensors, to determine how well somebody’s driving a car. So, you would install this app, and then the app can tell whether you do hard brakes, rapid acceleration, whether you’re speeding, whether you’re in an accident, so that was cool. That was interesting. Was a bit hard to find a business model. I think that was ultimately the problem that, “Okay, we have the technology. Now what?”

And I think that happens often in startups where it’s easy to hide your head behind the screen and code, but once you’re done and nobody clicks on your website, it’s much harder to know, “Okay, now what should we do?” So, this was cool for about the year, and then some recruiter contacted me for a consultant type of position at Bloomberg, at the Natural Language Processing Group. I thought it sound very interesting, so I went. Yeah, we hit it off. Then after a few months, they said, “Do you wanna join us full-time?” Yeah, I mean, I didn’t hesitate. I said, “Yeah, I’m down.”

WALLACE: Well, that’s a great story, though. A really natural trajectory. Could you describe for our listeners what is natural language processing? Some people might not know.

VIGAN: It’s an exciting field. Basically, we tackle any problem which come up in algorithms when you try to deal with natural language. In it’s most general form, I guess, what we like to do is we really want to send meaning, so the semantics of text. So, we want to teach a computer to understand a news article, for example. Really means, for example, now more concrete, one of the soft fields is named entity recognition, so that means, for example, you have a sentence, “I work at Bloomberg.”

So, now named entity recognition is the task of knowing that Bloomberg is an entity, so it’s some concrete instance in the real world, and then this ambiguation comes, and there’s this second step of named entity disambiguation were now there’s a person, Michael Bloomberg, there’s a company, Bloomberg L.P., and now you want to know that okay, he’s not working at Michael Bloomberg, but he’s working at Bloomberg L.P. This thing of really knowing the object of a sentence and linking it, for example, against the knowledge base. Wikipedia is a prominent example where, for example, if I wrote that sentence, you could underline it or link it through a hyperlink to the correct Wikipedia page.

Another task is language translation, Google Translate, for example. Question answering, and that’s one thing we work on or our partner team – the machine learning group – works on. So, you type a question, and you’re getting an answer. One of the cool things is called sentiment analysis. So, for example, you go on IMDB.com, you check out movie reviews, and you have an algorithm which tells you which reviews are positive about the movie, which ones are neutral, which ones are negative about the movie.

WALLACE: That’s so interesting. So, the applications of natural language processing for Bloomberg is dealing with data processing for its own analytics?

VIGAN: So, I can give you an example of the project I’m currently working on. So, this is, I would say, it’s called D-Duplication in Social Media. The fastest source of news is Twitter these days. A lot of our news guys are subscribed to Twitter streams, and that’s where they get their news from. Now, the problem is if a major event happens, there’s a lot of redundancy in the Twitter feed, so what I’m working on right now is to come up with algorithms which basically D-Dup tweets based on the semantics, based on the information they contain.

So, if you have a tweet which contains certain information, a new tweet comes in, might be written in a completely different way, but semantically similar, you would suppress this. So, here, the natural language processing helps internally. So, what’s a product of Bloomberg? There’s Bloomberg News, there’re a couple of TV shows, magazines, but the core product is the so-called Bloomberg Terminal, which is only present in the financial world. This is a desktop; you rent it on a subscriber model basis, and you have access to all these tools which you can use to make up your mind for investments, stuff like this.

And so, for example, what happens is we create, for example, headlines. We automatically create headlines for news stories which we extract from, for example, Quarterly, Quarterly Earnings, or we extract actual price of the stock from those reports, and this is all automatic using a machine learning natural language processing technology. So, these would be an example where actually our technology is not used in-house, but for the client. So, it’s both. It’s in-house and clients.

WALLACE: So, digesting and creating information in different formats, language based formats. That’s really, really interesting stuff. What’s a typical day like in the office at Bloomberg?

VIGAN: Yeah, so I would say it really depends on the phase of the project you’re in. So, for example right now, I’m with this D-Duplication task on Twitter, so I’m still kind of in the research prototype phase, so I’m reading a lot of papers these days and implementing algorithms which look promising. And then ultimately, when we converge to something, which then the business side also agrees on or thinks it’s useful for them, then we will polish the code. We have huge data volume, so everything has to be very efficient.

And then you deploy, and then as usual, things once in a while, something comes up and you have to maintain it, you have to maybe tweak it, you get some feedback from the business side, stuff like this. But I mean, in general, our team, we’re very research oriented, so we read a lot of papers, we go to conferences, we discuss, we have seminars. So, in that sense, it’s quite similar, actually, on a daily basis to what some PhD students’ day would look like maybe.

WALLACE: That’s interesting. And this is responsibility for the full life cycle of the product.

VIGAN: So, that’s cool because you learn a lot of different things. It’s not just like okay, now you’ve done this for five years, so now you know everything about natural language processing, but nothing about databases. But soon as your project involves a database, then you also need to learn about, “Okay, now how do I integrate the database? What’s a good way of modeling the data dependencies?” Things like this. So, yeah, I think you become better all around like a so-called foodstack programmer.

WALLACE: Interesting stuff. What is it that brought you to academia in the first place?

VIGAN: So, my Master was quite theoretical, so it was theoretical computer science, and I guess as a math student, you kind of scratch a bit on the surface of ongoing research, but you’re mostly focused on well-established textbook style of results. And I was always fascinated by the complexity of research papers. So, a couple of times, I had to present a paper in a seminar during my Master, and I always wondered, “Well, how do these guys come up with these proof technique out of the blue? They all must be geniuses.”

It was more like a challenge, a sports competition, and now you’re running 10 seconds, you want to run an 8.7 seconds. I was like, “I really want to understand this, and it looks fun.” I realized that there was a professor at CUNY, Peter Brass, who did work which I was really interested in in this more so-called discreet geometry, and I ended up there. I was like, “Okay, let me do one year. Let’s see how I like it.” Yeah, I liked it a lot, and I stayed.

WALLACE: And so, you talked also at the start about your being in the program, you were always somewhat ambivalent whether to stay in academia or to leave. Can you talk a bit more about that thought process?

VIGAN: I mean, for me personally, by the time I graduated, I was 34, so maybe a bit older than others because I used to work before I went to universities. So, most likely I would have done a post-doc in South Korea, one or two years, and then some tenure track, maybe three years, so by the time I would have a steady income, or let’s say a steady job, I would be almost 40. So, that was one thing which I wasn’t too happy about. Also, for me, it’s very important where I live. For example, I only applied to universities in New York City.

I mean, then, I thought about, “Okay, I wanna live in a cool city. I wanna do my type of research. Chances are pretty low that I’m gonna find this position, so I have to compromise one way or another.” So, either I move somewhere which wouldn’t be my first choice, and do what I wanna do, or I stay in New York and maybe I change my field. But still, if I’m lucky to find a professor position, yes. And we did this startup. I was very intrigued by it, and I also realized I think my skill set is more geared toward generality.

I mean, it sounds a bit funny saying this after a PhD, but I realized the older I’m getting, the more I wanna explore different areas, and if I took a professor position, it would be very focused, right? Okay, you publish this paper, you stay very focused. I mean, ultimately, I think with these big decisions, I always think, “You know, ultimately, it doesn’t really matter. At the end of your life, I mean, if you’re cool, you’re gonna have a good life in academia, you’re gonna have a good life outside of academia.” These big decisions, yeah, just make the best out of it. I think that’s what’s important.

WALLACE: And it sounds like also a gut feeling. You mentioned that working in the startup, you had a feeling of the pace of the work was something you enjoyed.

VIGAN: One thing, if you wanna prepare for working in tech or so, internships are definitely cool. I mean, definitely very useful if it’s a name which people recognize. Yeah, but doing your own startup, you’re just gonna learn a lot, and ultimately, even at Bloomberg, we’re very startuppy in the sense that we’re agile; this is a term that refers to how you do project management. We have this two-week spree. Although it’s a big company, we’re very startuppy.

WALLACE: So, what do you enjoy the most about your current work?

VIGAN: Well, I guess it’s really the fact that we implement academic findings, so results from papers, and we apply to our business side. This actually has impact on the business. So, it’s this nice bridge of bringing academic findings into industry or business environment. That’s pretty cool. So, you can still nerd out. It’s a pretty geeky floor we’re on. There’s a lot of logician, mathematicians, complexity theoreticians, machine learning people, and so on.

So, it has a very academic feeling, but at the same time, people are kind of aware of okay, I cannot just sit around here and think. At some point, there is also internal call centers. They pay for us to deliver some product to them, so you also have to be business conscious. I like this. I like this bridge, this intersection of academia and business, so that’s very rewarding, I think.

WALLACE: Yeah, the rubber hits the road, so at some point you all have to pull together, and be accountable, and be a team, and produce something, and make a difference. Are there other aspects that you like about it? It could be anything from lifestyle, company culture …

VIGAN: Yeah, I mean, it’s cool. It’s cool guys. We have fun together. I mean, I know it sounds stereotypical, but it’s actually true. We go out, we have beers. Even though we’re 16 people now in our team, yeah we go out. We laugh a lot during work. Some guys come in at 11:30 in the morning; some other guy comes in at 6:00 in the morning and leave at 3:00. Today I’m working from home because I had a delayed flight yesterday.

So, yeah. I mean, I guess in that sense, it’s similar to academia. Ultimately, nobody cares how many hours you spend in the library. It’s more about okay, where did you publish your paper? It’s similar here. I mean, like now, I’m working on this project and people are waiting for it. Yeah, I mean, I can take it easy one day. Nobody says something, but I mean, you just need to deliver.

WALLACE: Has finishing your PhD benefitted you in your career, having the degree?

VIGAN: For this specific position, definitely. I mean, it’s a requirement to have a PhD because it’s a research position, but even without, even if this wasn’t a requirement, I think what benefitted me for now having a PhD for this position is, especially since I’m reading quite a bit of papers, you’re getting used to the jargon. You can skim over papers and not only understand them, but also assess like, “Okay, is this now something trivial or is this something incredibly, incredibly novel?”

So, the PhD helped because now we have all this massive amount of tape. It doesn’t help you if you have the best algorithm, but it takes two years to train; nobody cares. So, you need to be very efficient. You need to write efficient algorithms, and that’s where the whole theoretical computer science aspect of my PhD comes in where you actually have provable bounds on the running time of your algorithms, and you can say, “Okay. We can apply this algorithm on this problem because by theory, it’s gonna terminate after two days or so.”

WALLACE: So, being able to apply theory to cases that have a lot of benefit to companies like Bloomberg?

VIGAN: Exactly, yes. Yes.

WALLACE: Maybe from your perspective now outside of the university, are there things that you wish you would’ve learned that could’ve prepared you for the work you’re doing?

VIGAN: Yeah, I mean, of course you can tailor your curriculum towards what companies want, and that’s easy to see. Just go on Glassdoor and see what companies are asking for. You can definitely do this, but I think ultimately, maybe you’re gonna regret it because those things, once you’re in the job, you’re gonna learn it anyway.

WALLACE: So, there’s a bigger picture there in terms of following your passions, and the kind of discoveries you can make when you’re following your passions may lead you to more enriching discoveries than just targeting what the employer says they need?

VIGAN: Yes, yes, yes.

WALLACE: That’s really good advice.

VIGAN: I think an internship is so important. At least one internship, yeah.

WALLACE: For the credential or for the student to get experience and really understand?

VIGAN: Actually, both. Actually, honestly, both. I mean, if you have a Google internship on your CV, that definitely outweighs your average grade or maybe even where you start the things like this, but also, I mean, to be honest, yeah, you learn a lot. You learn a lot. This summer, we had a couple of interns, and I mean, they change over three months. In the beginning, they didn’t really know how to use all these tools to manage code, some Linux stuff. I mean, comparably simple things, but it’s just annoying when you start somewhere new, and you don’t know these things.

Also for the employees, like, “Okay, now I need to spend three months teaching those things.” Maybe you joined a small startup. It’s a long time. You do an internship, you learn those things. You also learn what does it mean to be in the business environment. I think somethings also, especially if you spend a long time in academia, you might forget. I mean, a 5:00 meeting is a 5:00 meeting, and with the business afterwards you can go for beers and then laugh, whatever, but let’s be down for business right now.

WALLACE: What do you know now that you wish you would have known as a graduate student?

VIGAN: Enjoy the student time. I mean, enjoy the freedom you have with respect to schedules. You know, if you’re full-time – although I’m really happy everyday to go to work; I really mean it – you cannot just say on a Thursday afternoon, “Let’s go play tennis,” so really, sometimes you might think, “Oh, man. A PhD is so much work.” Of course, it’s true. I mean, somewhere I read the other cool thing about being a professor, you can decide which 70 hours of the week you wanna work; that’s a cool thing. That’s a nice thing to have. Other than that, I think be proactive about things.

I think when you do graduate studies, even maybe if you have deadlines by your professor, it’s usually him who reaches out to you, like, “Hey, what about the meeting? Do you have anything to say on this and that?” It’s much better the other way around. If you reach out to your team and say, “Hey, guys. I came across this really cool technology. Let me give a 20-minute presentation about it,” stuff like this. So, be proactive about things. It’s a very small effort, but it goes a long way. I mean, it just shows you can actually prioritize, you see what has to be done, and things like this. So, yeah, I think that’s something I just realized after working for a while.

WALLACE: No, that’s very good advice. I think our students can benefit from thinking about being proactive, taking responsibility.

VIGAN: Yeah, we’re constantly hiring. Careers.bloomberg.com and apply. Yeah, we have lots of jobs, and it’s a fun environment to be in.

(Music)

WALLACE: That’s a wrap for this episode of the Alumni Aloud podcast, coming to you from the Office of Career Planning & Professional Development at the Graduate Center. Be sure to stay tuned for more perspectives from alumni across fields who are working in academic and non-academic jobs. Also, be sure to look up our online resources at careerplan.commons.gc.cuny.edu, and like our Facebook page, or follow us on Twitter for updates from our office. See you next time and thanks for listening.

(Music)

Back to Top