DSI Student Lands Great Job at Facebook

Marika Lohmus
Marika Lohmus

Although she’s still a student, Marika Lohmus already has a great job.

After she finishes DSI’s master’s program in winter 2017, she’ll begin working as a risk project manager for Facebook. She interned at Facebook this summer, excelled at her work, and was offered a full-time position on its Risk & Payments team.

As risk project manager, she’ll help identify fraud and address financial or safety risks before the launch of any new products. The team is based in Austin, Texas, but has members in Menlo Park, California, Dublin, Ireland, and is expanding to Singapore.

Marika is a long way from home. She was born in Tallinn, Estonia, and immigrated to the United States with her family in 2001. Her father, a political economist, had taken a job at the International Monetary Fund, so the family settled in northern Virginia, where she attended school.

Growing up, she loved foreign affairs and government but also took an early interest in computer programming. At Cornell, her undergraduate school, she double majored in government and information science. After graduating she worked for Deloitte Consulting’s Analytics & Information Management practice, where she focused on data management and advanced analytics for major financial institutions. She also created large-scale, multi-year data strategies for executive officers. Eventually she decided to enroll at DSI, she says, to develop her machine-learning and data-visualization skills and become a data-science specialist.

In this interview, Marika talks about her passion for data science, her work for Facebook and her journey from Estonia to America.

************

Discuss the work you did during your internship.

As mentioned, I interned with Facebook over the summer, working with their Risk & Payments team, which is responsible for facilitating payments and for identifying fraud within their field of paid products. I mainly focused on creating models that would help identify individuals who may be using Facebook Marketplace to commit fraud and prevent their posts from surfacing on the platform.Half the team deals with any monetary transaction done on the site (payment for ads, purchases made in games, payment of developers, fundraisers, peer to peer payments, etc.) and the other team deals with investigating fraudulent transactions via these mediums. The feedback I received about my internship was very good - I was passionate about the project that I was working on so it was really easy to get involved. It was incredible to work on such an impactful project - every fraudster I caught meant that there was at least one less person cheated out of his or her money.

Can you give examples of how you prevented fraud during your internship?

Facebook Marketplace is a bit of an anomaly - even though there are no payments made directly over Facebook (typically you agree on a price, meet somewhere and exchange goods and money offline), the risk team wants to ensure the trust and safety of Marketplace users. Many of the types of scams we see have been around for a long time, except that they have now been adopted to the Facebook platform.

  • One case is the 'cheap car scam' - they put an image up of a nice looking car for only $2000. When you message them about it, they tell you that the car belongs to their friend / sister / mother-in-law and they are out of the country - you should email them at some generic email. There they ask you to buy pre-paid cards for $2,000 and to send them the security numbers, after which they will ship the car right to you for free! Of course, they disappear as soon as they get the money.

  • Another case could be a person that re-sells the same phone over and over again. They get paid via PayPal, promise to ship it to you (may even show a picture of a shipping receipt), and then block you as soon as you start wondering where your phone is.

These types of fraud are done by people on fake accounts (false profile created just for committing fraud), compromised accounts (real accounts that were 'hacked' by a bad actor), or via the fraudster's real accounts. Such fake accounts are against Facebook policies, and they have a dedicated team for identifying and preventing inauthentic accounts from surfacing on the platform.

Over the summer, as an intern I developed models that would score a person's account and their marketplace post on how suspicious it looked before the post went live. Based on a confidence metric, we could ban the person from Marketplace, submit the post for manual review, or let the post through.

Can you elaborate on why your family came to America?

My family moved from Estonia so my father could work for the International Monetary Fund and my mom went to Georgetown for her master's degree. After graduating, she went to work for the World Bank. They are both incredibly brilliant in economics and public policy. My dad actually graduated from Columbia’s SIPA (School of International and Public Affairs) back in '94! They have both inspired me in so many ways, especially by encouraging curiosity about the world - we traveled a lot when I was young.

How did you get interested in programming?

Thomas Jefferson High School, which was ranked first in the nation while I was a student, was where I found my passion for programming. I started coding when I was 11, but always did it on an ad-hoc basis. Getting a formal education in the fundamentals of programming in high school, though, was huge. In addition to introductory programming, I took AP computer science and artificial-intelligence courses. My senior project was an agent-based model trying to model population movements in response to global warming. I don't think I would have had the opportunity to take such classes elsewhere.

Was your interest in computing born from your love of games?

There's this old children's website called Neopets.com, where you get virtual pets that you feed with items bought using virtual currency called neopoints that are acquired by playing flash games. In order to get more neopoints, I started writing scripts that would automate, for example, buying rare items at a cheap price and selling them at a much higher one. I'd also write image-recognition scripts that would play these flash games for me. As Neopets improved their fraud-recognition systems, I'd constantly get banned but would figure out how to get back online and for years I beat their systems. It's one of the reasons I thought working for the Facebook’s Risk Team would be perfect for me - I'm actually doing the opposite work from what I had done as a child!

Leaving Estonia for America: Was that hard on you?

Coming from Estonia was definitely tough, though my dad insisted I take private English lessons to get me ready for the trip - I wasn't that good. I struggled through the first year of school but became fairly fluent by the next year. My parents were also strict about only speaking Estonian at home so I wouldn't forget it - though I have developed an accent and forget words here and there, I think I've done pretty well for having lived 16 years abroad!

And the best feeling is randomly meeting other Estonians in the city. Usually I'll hear an Estonian accent or overhear a few words to find people, but Estonians also have a very distinctive look to them. I saw Jaan Altosaar, a Ph.D. student, in the DSI lounge once and thought “he looks so Estonian.” I called my father to see if Jaan would recognize the language - and he did! He mentioned he knew a few other Estonians at Columbia so we had a little get-together afterwards.

In terms of gender, what has been your experience working in the tech world?

I never really thought it was weird being a woman in technology until I started my job in tech consulting. I recall sitting around a large table during my third month on the job and realizing that out of the 20-some folks sitting there discussing a large implementation problem, I was one of two women. It was definitely tough to prove that I knew what I was talking about and it took a few weeks on every project to get people to take me seriously.

It's definitely an issue - who wants to stay at a job where they do not feel comfortable, or always feel like they need to work harder to prove themselves than their peers? It's incredibly important for women to stand up for other women in fields where we are underrepresented. My most influential relationship was with a female senior manager who could just control the room the moment she walked in, and would not stand for being cut off or treated any different than the other leads.

You have prominent women in your family. Did they influence you?

Estonia is not the best example of gender equality. I believe we still have the highest gender pay gap in Europe, and the politicians that make up the parliament are largely male - though we do have our first female president! However, my family has an incredible track record of influential women that is quite shocking for Estonia - my grandmother was an industrial engineer and eventually worked for the Ministry of Finance and put together the first budget for the newly independent country in the early 1990s. My aunt is an influential pastor at the Lutheran church and recently bravely spoke out on behalf of gay rights. By the age of 29, my mom was the Deputy Secretary-General at the Ministry of Finance, a highly respectable position in Estonia. She is the hardest-working woman I know, and I don't think I would have had the drive I do without her.

Will you use data science at your job?

For sure. Even though I am not in a 'data scientist' role, my Facebook team builds models to assess the risks associated with every transaction and encourages everyone to keep learning and to apply the latest and best models to improve both precision and recall.

It’s your last semester at DSI. Have you enjoyed the master’s program?

Coming to Columbia and to DSI was definitely the right move for my career. I always loved playing with data outside of work. For example, I'd build a model to fill out my NCAA March Madness basketball brackets every year - and naturally win. And when I lived in Pittsburgh, I used a predictive model to guess how many protesters would be outside of the Pittsburgh Planned Parenthood office during my daily walk to work. But I am so happy now to be getting a formal education in data science and I’m excited to use these skills to help improve people’s lives.

--Robert Florida



550 W. 120th St., Northwest Corner 1401, New York, NY 10027    212-854-5660
©2017 Columbia University