Top left to right: Yumeng Ma (University of Michigan), Steven Yin (Columbia), Michael Spece, Carnegie Mellon University), Quinn Donahoe (University of Pittsburgh), Cristian Gavrus (UC Berkeley)
Bottom left to right: Ai Xu (Harvard), Elyse Norton (NYU), Aybike Ulusan (Northeastern),Yuija Zhou (Boston University)
It’s safe to say that Wayfair’s Data Science team has a decidedly academic bent. After all, nearly half of our 83 data scientists have PhDs. Their backgrounds vary–from nuclear physics to economics, neuroscience to biochemistry–but the experimental mindset, analytical rigor, and deep problem solving skills that these team members developed in their graduate programs make them all excellent data scientists.
This kind of academic crossover hire is becoming more and more common within the tech field, but an experience gap still exists between academia and industry. Without hands-on exposure to the kind of work you can do with mathematical modeling and coding outside of an academic context, many students in quantitative PhD programs are not aware of the opportunities available to them in the field of Data Science, or do not have the confidence to pursue them.
Wayfair’s Data Science Immersion Program was created as one way to bridge this gap, and we are excited to announce the completion of its second round! From June 18th-22nd, Wayfair again hosted a group of 9 promising graduate students from a variety of subjects and universities to sharpen their data analysis chops and give them a taste of what it’s like to work as a data scientist at a large ecommerce company. “The difference between a quantitative PhD and a data scientist is the training,” Dan Wulin (Director, Wayfair Algorithms & Data Science) told participants on their first day, “and this training can be done quickly.” Over the course of the week Wayfair aimed to jumpstart that training process.
Getting the Lay of the Land
First, in order to give a broad view of the possibilities of the workplace, each day participants spent an hour or two in Deep Dive discussions with Wayfair data scientists who exposed them to current projects underway in personalization, pricing, marketing, and merchandising.
“I liked the Deep Dive presentations,” Steven Yin (PhD Student, Industrial Engineering, Columbia) remarked when asked what was the most valuable portion of the week. “They gave me a very concrete idea of the kind of projects that people are working on at Wayfair.” Elyse Norton (PhD Student, Cognition and Perception, NYU) likewise commented that, “Prior to the immersion program I felt like I had a vague understanding of what it meant to be a data scientist and as a result was unsure about what companies were looking for in a data scientist. Now I feel like I have a much clearer picture that will help me moving forward in my job search.”
To round out the Wayfair tech experience, participants were also given a personal tour of what Wayfair Next (the company’s VR/AR wing) has to offer, and were given tips on how to prepare for data science case interviews.
Getting Their Hands Dirty
But these talks were only half of the equation. Participants were also divided into teams which were each given a live data set and a problem to solve by the end of the week–predicting gross revenue margins, developing fraud detection, or improving image classification. Following up on feedback from last session, less time was given to instruction or presentations and more was allocated to “getting your hands dirty” with the data, as Zhenyu Lai (Associate Director of Algorithms, Wayfair) put it. And the experience paid off.
Yuija Zhou (PhD, Mathematics, Boston University), whose team worked on developing a method of fraud detection, for example, noted that “The opportunity to work on real data was a very valuable experience. Not only did I learn a lot of technical knowledge, I also experienced and participated in the kinds of conversations that happen in a real workplace.” Her team member Cristian Gavrus (Postdoc, Mathematics, UC Berkeley) further elaborated on the value of the time crunch the boot camp imposed. “The biggest difference between working on academic vs. industry products seemed to be the pace at which we have to deliver results. Specifically, given the limited time, we were encouraged to come up with results fast and then improve on them later if needed.” Luckily a hackathon on Thursday night and a group of on-hand data scientist mentors were in place to speed along the process.
Getting Some Tips
Wayfair data scientists took time out of their schedules to answer participants’ questions, troubleshoot, and listen to practice presentations. This arrangement proved beneficial to both sides, allowing participants to “see first hand how data scientists at Wayfair solve problems” (Steven Yin PhD Student, Industrial Engineering, Columbia), and providing mentors with valuable practice in management.
As Zephy McKanna, mentor of the fraud team, expressed, “I think this was definitely helpful in terms of management skills, particularly with regard to hiring. One of the things a manager (around here) has to do is determine whether they’ll be able to work with someone in a very short time (one or two interview slots of ~45 minutes each. I think this really helped me to focus on the things that matter most to me, such as communication style, clarity of thinking, and the balance between correctness (‘here is the answer you expect and will probably work’) and creativity (‘here is something that you haven’t thought of, and probably won’t work, but if it does: wow’).” Fellow mentor Frank Ma added that beyond offering insight into the work and thought his own managers put into project oversight, the mentorship experience had the added benefit of giving him the opportunity to work more closely with colleagues outside of his core group.
At the end of the week participants presented their findings and fielded questions from an audience full of members of Wayfair’s data science and algorithms teams. Yuija Zhou (PhD Student, Mathematics, Boston university) commented on the usefulness of these questions in orienting her to the concerns of the business sector. “The mindset is very different,” she stated, “Questions asked in industry are business driven, whereas they are curiosity driven in academia.”
We have been impressed with the caliber of participants in our programs thus far, and look forward to continuing the program in the future. Director of Data Science at Wayfair Dan Wulin recently reinforced this idea, stating that “The Immersion program is a very important piece of the Wayfair Data Science recruitment strategy. We’ve had a ton of success at hiring data scientists out of a variety of Ph.D programs and Immersion lets us form relationships with top candidates before they enter the job market. The program is a two-way street, since we’re able to engage the candidates with Wayfair and they get an opportunity to learn more about data science in industry by working closely with practitioners for a week in our Boston office.”
We are looking forward to welcoming our next Immersion cohort in January 2019. Applications can be found here!