The Core Data Science (CDS) team at Facebook works with a variety of teams on an ever-changing range of projects and products to advance the frontier of data science. Because of the nature of the work done at Core Data Science, many of the team’s research scientists come from a broad range of academic backgrounds, from the social sciences to other STEM fields. “It’s a common misconception that you need a PhD in computer science to do research at Facebook,” CDS Research Scientist Manager Aude Hofleitner explains. “But for teams like CDS, diverse perspectives enrich and expand our work.”
Hear from four research scientists on the CDS team about their non-computer-science paths to Facebook, what inspired them to make the transition from academia to industry, and what they’re working on today.
Utilizing statistics to help identify inauthentic content for removal
Yilin Zhang earned her PhD in statistics with a minor in computer science from the University of Wisconsin-Madison. Her research interests include applied machine learning models, social network analysis, spectral clustering, community detection, and network-driven sampling. Yilin loves being able to apply her personal interests in statistics to her work at Facebook. During her interview for the CDS team, she had the opportunity to present one of her PhD projects related to the work she would be doing. And more recently, she’s been applying her knowledge in statistics to improve data distribution for the team in charge of removing inauthentic content on Facebook’s platforms. “The CDS team is unique because we don’t just focus on publications or work on one specific product,” Yilin shares. “You get to collaborate with a lot of different teams and work across a variety of products. There’s so much opportunity to explore new fields of work and expand upon what you’ve studied or worked on.”
While supporting teams over the past two years, Yilin says she’s also been able to improve both her communication and coding skills. “I came to Facebook with an understanding of R, which I used while I was studying, but I’ve become so much better at coding in Python during my time here,” she shares. “I remember preparing for my technical interview and anticipating how I would handle the coding questions. When I look back now, I can see how much I’ve learned.”
Outside of work, Yilin is a member of Facebook’s Women@ Resource Group and loves taking part in the organization’s social activities to encourage more representation among women in the industry. Her advice for women looking to join the CDS team? Be open-minded and just apply. She offers, “When I read some of the research and papers coming out of Facebook, I didn’t think I had the right experience. But after going through the interview process and speaking to various people on the team, I realized there was a place for me here.”
Expanding Data for Good research with a background in economics
Monica Bhole has a PhD in economics from Stanford University. Before joining the CDS team at Facebook, she worked at Pandora on advertising, experimentation, and economic mobility. Her areas of focus include causal inference, experimentation, and understanding the impact of the internet and social networks on economic mobility.
While studying, Monica didn’t think about working in technology. “I was initially interested in academia or working in policy,” she remembers. “It wasn’t until I spoke to a few PhD graduates that I learned this was an option for me.” Monica had been primarily familiar with R from her PhD work and Python from her previous job when she was hired. During Facebook’s onboarding bootcamp, she was able to build upon her existing coding skills and learn more about engineering at Facebook. “Bootcamp helped me become comfortable and familiar with Facebook’s systems. It was engineering-heavy, and I learned a lot.”
“Working on the CDS team is a rare opportunity to collaborate with social scientists while also informing product changes through engineering,” Monica beams. “I love the flexibility to be able to do research and work on the product. It keeps things exciting.”
Monica’s research is often guided by her background in economics. One of her most exciting projects is her work to develop commuting zones, which are part of Facebook’s Data for Good program. “When doing economic research, we will often conduct our analyses using commuting zones, which are incredibly valuable,” she explains. “But the challenge is they’re limited to U.S. data, so if you want to look at data from other parts of the world, you’re stuck. This is the problem I wanted to solve, and I realized it could really push research forward not only internally, but also externally through the Data for Good program. I love having the freedom to shape my own research and road map, all while making a strong impact in the community.”
Drawing on an expertise in political science to study elections
Pablo Barberá earned his PhD in political science from New York University and is also affiliated with the University of Southern California as an associate professor. Before Facebook, he served as an assistant professor of computational social science at the London School of Economics and Political Science. His research centers on computational social science, natural language processing, social networks, and political behavior.
While Pablo was at USC, his research was focused on understanding Facebook’s impact in the political process. Soon after, he was offered a position to join the company full-time. “Working at Facebook gives me the opportunity to work on all the things I enjoy about academic research, but with more flexibility,” Pablo explains. “In academia, there is often a pressure to ‘publish or perish,’ so I found myself thinking about work all the time, even on the weekends. But in industry, and at Facebook Core Data Science specifically, that pressure isn’t as pronounced, so I feel comfortable leaving work at work.”
Today, Pablo spends the majority of his time coding and working in a more technical capacity—by choice. “While I already had some experience with coding prior to joining Facebook, going through Facebook’s mandatory six-week Bootcamp program was eye-opening and really helpful in enabling me to pick up new skills.”
Even more, Pablo says, he appreciates having the ability to choose projects that interest him and having ownership of his work. “My team has the autonomy to complete work that has an impact far beyond Facebook,” he explains. “We’re currently working on the largest election study that has ever been done, which I find extremely exciting. Most companies aren’t driven by social science research, and there’s a big disconnect between industry research and academia. At Facebook, we’re bridging that gap.”
Making a meaningful impact with a political science background
Thomas Leeper has a PhD in political science from Northwestern University. He was an associate professor in political behavior at the London School of Economics and Political Science for three years before making the transition to industry. In 2018, he decided to join the CDS team at Facebook, where his research focuses on experimentation, causal inference, and computational social science.
Thomas recalls the time he spent brushing up on the basics in preparation for his Facebook interview. “You can code in a language of your choice, but if you’ve never coded in front of anyone or had to explain your thought process, it might be intimidating,” he says. “It’s a good idea to study some basic algorithms or take an introductory course before the technical interview.”
He continues, “While more than half of our CDS team members don’t have degrees in computer science, we consider ourselves an engineering team. You see a lot of papers published by the CDS team, but there’s much more going on behind the scenes. Our day-to-day work means we all have to be excited to do software engineering, and CDS provides the tools, resources, and language to do it.”
Thomas says that when he first joined CDS, he was struck by how quickly he could make an impact — without a computer science background. “When I was interviewing at Facebook, I had recently published a paper that took me six or seven years to complete. By contrast, I’ve been at Facebook for almost three years, and I’ve completed multiple projects in a timeframe, scale, and scope I never would have been able to achieve before. Our industry is dynamic, and we’re solving new challenges all the time. The impact I could make in an academic setting could never translate to real-world differences the way it does at Facebook.”
Interested in learning more about the CDS team? Check out their research team page.