Two more of our PhD students were recently awarded a prestigious spot in the postdoctoral training program called “Insight.” This program involves summer intensive technical training to bridge academic and scientific methods to industry application, along with subsequent placement in a major data science / industry context.
Dr. Bryan Kerster and Dr. Till Bergmann finished their CIS dissertations in 2016, and have since moved on to exciting industry positions after their Insight activities. Bryan is a data scientist at Facebook, and Till is a data scientist at SalesForceIQ. Both are working with massive natural data, and applying their skills (and acquiring new skills) of coding and infrastructure for working with these data.
We asked them about their experiences.
What did you like about Insight, and how did CogSci training help you?
Bryan notes: “Getting an inside look into the roles at dozens of companies allowed me to gauge which kinds of roles in data science would best fit my interests and strengths. I think there are very few other ways to learn about the state of data science in such a short period of time,” and observes that
…my training in Cognitive Science was extremely helpful. My research already had me using many of the tools frequently used in industry which allowed me to spend more time focused on the details of my project, and less time on learning basic tooling. Additionally, during the latter half of the program which focuses on working with other fellows to prepare for interviews it started to become clear how much CogSci and data science overlap, especially in topics like experimental design, statistics, and probability.
Till makes some of the same observations about Insight, its intensity and exciting connections to industry, especially learning the term “data scientist” has varied meanings, and that pairing with an industry position can be subtle, based on your specific interests, skills and their need. Regarding the importance of his work in cognitive science, he notes that:
My research in Cognitive Science was very applied to begin with, and many tools I used during my PhD were directly applicable at Insight and generally in industry, ranging from NLP to controlled experiments (A/B tests, as they are known in industry). I realized that compared to some other PhD students, we behavioral scientists know a lot more about experiments and how to run them. A simple A/B test is valued by companies a lot more than you would think coming from CogSci! Like Bryan, I was pretty comfortable with my technical background and could thus focus on my weaknesses such as understanding the business side and prepping for interviews.
What advice could you share to graduate or undergraduate students interested in data science?
Bryan and Till both advise undergraduates to develop technical practices — direct experiences working with real data, going beyond abstract exercises from classes. Bryan notes that “the skills that are going to most important for an entry level data position are going to be around experimental methods and statistics. Make sure to learn those topics very well.” Till shares a similar point: “It’s really important that you can show that you did something actively!”
For graduate students, Till says that making the right connections between your work and data science will ensure you can talk in engaging ways in interviews. Cognitive theory is very important in our scientific field, but expanding the relevance of your own work to something that interviewers from totally different areas will find valuable and relevant is challenging but quite critical. Bryan notes that cognitive science is a great place to make these connections, but also highlights that gaining expertise in other domains, particularly machine learning, is absolutely critical.
Congratulations to Dr. Kerster and Dr. Bergmann!