Alumni Spotlight on Sam Swift: How a Degree in Social Science Can Lead to a Career as a Data Scientist

I am a social scientist with a background in decision-making and behavioral economics at Carnegie Mellon University.
I was part of the first undergraduate class in Decision Science, which combined the psychological and economic perspectives on human behavior. I went back to CMU to complete a PhD in Organizational Behavior at the Tepper School of Business, where I conducted lab and field research on decision-making and negotiation.

Along the way, I’ve developed expertise in software development and statistical programming with R. I spent three years as a developer in a software consulting startup and two years as a postdoc managing a team of statisticians and developers. The fact that I was passionate about behavioral economics and decision-making, but also really enjoy working in a software development environment, contributed to my decision to ultimately pursue new opportunities outside of academia.

Currently, I am a data scientist at Betterment, the largest and fastest-growing* automated investing service, where I am responsible for insights into and analyses of our customer’s investing behavior.

Coming out of academia, one of the biggest challenges was understanding my own preferences and relative strengths in industry. In six weeks at The Data Incubator, I was able to get clarity on the questions it would have taken me years to develop on my own. I realized that I wanted to join a mid-size startup like Betterment, rather than a nascent founding-phase company or established corporation. I was not as interested in building out basic data infrastructure, but I did want to join a team on which I could quickly be very influential.

I also realized that there are many data scientist positions for which I am not a great fit, but that matching is a two-way street. Rejection feels like a big setback but is likely an indication that the job would not have been fun and fulfilling. The interviews that went poorly were actually helpful and informative because they were likely early signals that I would not have enjoyed the position going forward. In my case, the interview with Betterment stood out as a great fit on both culture and skill requirements, and they agreed.

The intense incubator experience was also a great way to quickly transition my thinking and language from academic abstraction to business pragmatism. Like miscommunication between any two fields, I found that there was lots of common ground on ideas, but that it was obfuscated by specialized jargon on both sides.

There is also a learning process around the tradeoff between creating something useful and creating something perfect. The business environment demands more emphasis on the former, which is a lesson I’ve learned more than once. The Data Incubator gave me a great head start on those transitions, allowing me to focus on more important substantive ideas during interviews and my first weeks on the job.

The Data Incubator is an opportunity to find the best next step in your career, but it is not an automatic process. In my academic roles, I got used to well-defined programs with a diploma reliably granted after all of the steps had been accomplished. The most exciting opportunities in industry are not always well defined, and companies may not even know they are looking for you. Finding the right fit depends on the realization of what kinds of fields and roles you are interested in, and what your specific background prepares you to contribute. It’s not too early to start that process while you are applying and before you start the program.

My social science background was essential to feeling like Betterment was the right fit. Betterment is a disruptive financial services startup, in large part because of its emphasis on investor behavior and data-driven insights. The company already had a great team of data engineers and analysts, but Betterment wanted to add to those technical abilities to gain a deeper understanding of user behavior. I use my social science training every day as we generate new hypotheses and analysis plans to improve our product and help customers become successful investors. Whether it’s R, Python, SQL, MapReduce or any other tool, the tools are just means to an end; the real purpose of data science is to ask and answer interesting, useful questions about people and business—exactly the contributions social scientists are prepared to deliver.