NYU Center for Data Science Incredible Alumni: Alex Simonoff

Data science is a rapidly growing area of study. Yet many people are still left asking, “What actually is a data scientist?” If there were a simple answer, data science wouldn’t be all the rage it is now. What makes data science so unique, among other things, is the application of the knowledge and skills it takes to be a data scientist to a vast, interdisciplinary array of studies and professions.

Data is everywhere, and data scientists are following. CDS provides students with tools and networks to set them up for success in whichever industry they enter. But don’t take our word for it… Allow us to introduce one of NYU Center for Data Science’s Incredible Alumni: Alex Simonoff!

Alex graduated from the CDS Masters in Data Science program in 2018 and works now as a Data Scientist at Spotify. She spoke with us about interview tips, imposter syndrome, and her current work at Spotify.

Photo. courtesy of Alex Simonoff

This interview has been lightly edited for clarity.

What drove you to pursue a Masters of Data Science?

I went into actuarial science, then data science and insurance, and from there I found my way to CDS. When I finished undergrad, the sexiest job in the world was an actuary, so that was my initial foray into the workforce. I was at AIG at the time, and we had a science department. If you name something science you’re just begging for people to be intrigued, so obviously I was intrigued!

I interviewed for the role and realized there was a lot more to data science than I had any idea about. I got the job, and after about six months there I took my GRE and applied to the Masters of Data Science program because I wanted to apply myself more. There were a few people at work who were actually doing part or full time Masters at CDS, a gentleman named Yoon Kim was in my department and recommended it.

Why did you choose CDS?

At the time I was applying, my choices for Masters of Data Science programs were essentially Columbia, which had graduated one class, and NYU, which had graduated two or three. It definitely helped that NYU was already my Alma Mater, so I knew what I was getting into, and my dad is actually a Professor of Statistics here. I grew up 45 minutes away on Long Island and of course, if you have family as faculty, you get a pretty great discount for an undergrad degree. I also looked at some CDS graduates and faculty and they were doing things I thought I would love to do one day. So I applied to both, and when I got into NYU I immediately withdrew my application from Columbia.

Did your family background have an influence in your course of study and career path?

Yes, for sure. It’s been really cool to have that partnership with my dad. I studied math in undergrad, and once I took his statistics class I was like, ‘Oh, actually, generalized linear models are really cool and fun!’ And he was like ‘Ohhhh yes, yes they are!’ He says he brings me up constantly in his class because I do what a lot of his students would love to do one day. I went to the Joint Statistics meeting for the first time as an attendee, but when I was a kid, my dad would go to speak and bring me as a guest, take me to all the tables with the free swag. And this time I was like, ‘Oh there’s a talk by Netflix on A/B testing, I should probably go to that.’ Any time I visit my parents, we’ll talk about what I’m working on, without disclosing anything you would have to sign an NDA for, I’ll tell him about the challenges I’m facing, and he’ll actually give me some recommendations on things to try.

What can you tell me about your job, without handing me an NDA?

I’m at Spotify, and we have so many different data science jobs. Machine learning engineers essentially build production codes to do recommendation systems and multi-arm banded testing, things you don’t necessarily think an engineer with a software engineering background would be able to do. Research scientists are more PhD-level computer or data scientists, or mathematicians who research neural nets.

Neither of those really appealed to me, so I followed the traditional product insight/data science for product analytics route. My day entails a good deal of meetings to explain things I’ve found in data. I run a lot of A/B tests to make sure we are making the right changes to our product. Once an experiment is done, we measure differences in metrics to see if we’ve actually improved our targeted metric. I also build dashboards to track business metrics over time. It’s all about quantifying the value brought to the company by the product I support and innovating it in ways that make it even more impactful.

Has having a Masters in Data Science degree from CDS helped you in your career?

Yes. I wholeheartedly believe the only reason I now have a job in tech is because of CDS. I did the program part-time for three and a half years and in my first year in the program I felt wasn’t getting the opportunities I wanted in financial data science, so I interviewed with companies from completely different industries.

Etsy called out to me as this little grass roots tech business staffed with people who cared deeply about their jobs and their place in helping the world of creators, it was this beautiful escape from all the things I didn’t like about tech and my current job at the time. During the interview they made it very clear they were excited about my part time study at CDS, and that was one of the main reasons they had extended me an offer.

When I started the program I had been a data scientist for about a year, and most of the work I was doing was presenting decks based off of insights that other data scientists had come up with because they didn’t want to deal with decks and stakeholders. Once I was able to ask them to explain their logic behind something like using one generalization technique over another, I had a seat at the table.

You mentioned you completed the program in three and a half years. What was that like?

It was definitely not easy, but I didn’t sign up for easy. I like to stay busy, I have a thousand hobbies, so school became one of my hobbies. I steadily took two courses each semester for a while, but I was interviewing for jobs when it came time for a capstone project, so I pulled back and discovered the joy that is taking one class! That’s why it ended up being three and a half years. Taking two classes at a time is a totally reasonable workload if they’re the right classes. I was very transparent with Loraine and the other administrators, and everyone was accommodating.

Were there any classes you took in the program that stood out to you?

During my Masters I realized I don’t want to do machine learning. However, Rosenberg’s Machine Learning class was the most fun class I ever took! I learned ultimately that I didn’t want to be doing the super technical coding aspects of data science, but he made it so much fun that I was always excited to go, and a lot of the material in that class comes up in interviews.

I took a bunch of classes in the computer science department, and that’s been immensely helpful in getting me up to speed with my engineers. I feel I’m able to talk to them person to person even though they have more knowledge about databases and I only have one course worth of knowledge (the course was a great primer). My other favorite class was a course by Brendan Lake and Todd Gureckis from the psychology department about neural nets and reinforcement learning and how it plays out in human behavior and learning.

As a woman in Data Science, has your experience in the industry been affected by your gender?

Gender is a very strange way to experience different things, and I do find it’s harder to be taken seriously in some organizations. Luckily the last two I’ve worked at have been around 50/50 male/female and other genders (there have been non-binary folks as well). I find that teams work so much more efficiently when you have that balance because of the diversity of thought.

I’ve been on panels before talking about my experience in data science and I’m the one woman out of four men, I have been the only woman on a team of six engineers for a product, and it’s not always easy to get a word in. I’ve had to step out of my comfort zone with being assertive and making sure that I’m getting the treatment and opportunities that I want. Ultimately at my first job I wasn’t getting those opportunities and I do think gender might have had a role in that. Luckily since then, it’s been a little easier.

What advice would you offer to young women getting started in data science?

One thing women often do is undersell themselves. They say, “I have to be 100% confident about something- I have to know the ins and outs of a neural network before I put that on my resume,” whereas I’ve seen some folks that aren’t women say, “Oh yeah, I’m really good at Java, I’m really good at Python, I know all the stuff in SQL,” and then when you start asking questions, that falls apart.

I’ve been in interviews with women where they’ve said something like, “I implemented this really cool model and got this really cool result,” and I’m like, “Well, you should really put that on your resume.” Never be afraid to own the skills that you have and the work you’ve done because at the end of the day, your confidence is going to take you so much further than questioning your skills.

Regarding interviews, do you have any advice for first time interviewees?

Interviewing is very hard. There are questions you can’t prepare for, but I’ve picked up tips, especially now that I am interviewing people myself.

The first thing I tell everybody is to talk about what you are doing. We have a technical screen where we ask a basic coding question requiring candidates write a function or something like that, and time after time people don’t talk it through. It’ll be complete silence for about five minutes, they’ll write something down and go, “Yeah, that’s it.” It doesn’t give me the opportunity to feed back to them when they might be going down the wrong path. If I can see someone course correct, that tells me a lot more than if they just got it wrong.

Also, I never stop interviewing. A lot of people just assume that since I have a job at Spotify I’m a great interviewer, but I’m not. Which is why I work on it constantly and which is why I’ve gotten okay at it, but it took me a while to get a job in tech and it was ultimately starting a Masters program that propelled me. Even if you get a job, don’t necessarily just sit at that job, even if it’s the best job you’ve ever had. I fully believe I’m at the best job I’ve ever had, but I still interview at least once a year because it’s a muscle you have to flex to make sure you can use it when it actually matters, like when you get the interview of your dreams.

What has been the most rewarding moment of your career to this point?

Getting my current job! Growing up I was a musician, I’ve been super interested in working in music and working for a company where I care about the outcome. When I apply somewhere, I look at all the apps I use on my phone and put applications in at all those places. Spotify was the big one I was crossing my fingers for, so getting an offer from them was the most fulfilling moment. It felt like I’d been doing something right! That I had the ability to do this, even without a Masters degree at the time, the whole imposter syndrome silenced for a little bit.

How do you deal with imposter syndrome?

Pretty much everyone I’ve connected with on a personal level at work has expressed some degree of imposter syndrome. Sometimes I remind myself that the 80/20 rule is a very real thing, you can get 80 percent of the way with 20 percent of the work. I’ll usually aim for 90% of the way, but at the end of the day, if I try to perfect every single report I put out, it’s not going to be as impactful as doing interesting work and being confident in myself and what I’m putting forth.

I recently completed my first ever causal inference deep dive, and I was like, “Oh, I don’t think I controlled for all of the confounding factors.” So I made that a note! I didn’t say, “I’m going to sit in front of this for another three weeks and try to control for every confounding variable,” because I never would. If the work tells a story and helps us move forward, then I’m doing my job and I’m doing it well.

By Mary Oliver

Official account of the Center for Data Science at NYU, home of the Undergraduate, Master’s, and Ph.D. programs in Data Science.