Interview: Brad Klingenberg, StitchFix on Decoding Fashion through Analytics and ML

We discuss the challenges in making personal styling recommendations, unexpected insights, interesting trends, motivation, advice, desired qualities in data scientists and more.

brad-klingenbergBrad Klingenberg is the Director of Styling Algorithms at Stitch Fix in San Francisco. His team uses data and algorithms to improve the selection of merchandise sent to clients. Prior to joining Stitch Fix Brad worked with data and predictive analytics at financial and technology companies. He studied applied mathematics at the University of Colorado at Boulder and earned his PhD in Statistics at Stanford University in 2012.

First part of interview

Here is second and last part of my interview with him:

Anmol Rajpurohit: Q4. What are the most underrated challenges of using Big Data and Machine Learning for personal styling recommendations?

Brad Klingenberg: Fashion is complicated. To make our recommendations we use data about our clients, our inventory and the fashionhistorical data from all past fixes. A lot of this is data that we create ourselves. For example, when our merchandising team buys new inventory they create structured data describing the items in great detail. This isn’t easy and requires the time of fashion experts. Getting the data right is very important, as is understanding the different factors that contribute to whether a client will love her fix. It’s not enough for clothes to fit - they also need to match a client’s style and price preferences and complement her current wardrobe.

AR: Q5. What have been some of the most unexpected insights earned through applying Analytics for personal styling?

customer-feedbackBK: The amazing effectiveness of combining humans and machines continues to excite me. The key to this approach is finding the strengths of each. Learning where the boundary should be and how to optimally use feedback is an extremely interesting problem - I am sure we will see much more of it in the coming years.

Another important aspect of our business is how closely our clients are aligned with us. It may sound obvious but it’s actually quite important. Our clients want us to do a good job, and they want us to get better with feedback. This encourages them be effusive and thoughtful with their feedback - a critical source of data for all parts of our business.

AR: Q6. Which of the current trends in Big Data or Machine Learning are of great interest to you?

ipython-notebookBK: I’m most at home in Python, but am excited by the ever improving ecosystem of data science resources. Tools like the iPython notebook and pandas in Python and dplyr in R are making it easier and easier to manipulate and quickly visualize data. The trend toward easy reproducibility is important in industry, and certainly for academia as well.

On the statistical side, I enjoy following the continued development of tools for getting the most out of sparse data. I’ve also found myself very interested in some more classical statistical tools, such as random and mixed effects models.

AR: Q7. What motivated you to pursue a career in Data Science?

statisticsBK: I studied applied math as an undergraduate and found that I was always the most interested in projects that involved data. This led me to study statistics in graduate school. I did several internships during the summers and found that I loved the practice of applying statistics and machine learning to solve business problems. I was lucky to graduate at a time when statisticians and data scientists were taking on an ever larger role in industry.

AR: Q8. What is the best advice you have got in your career?

core-businessBK: For statisticians and data scientists especially I think the best advice is to stay close to the core business of the company they work for. This makes it more likely that you’ll be working on the projects that are important to the company - this is generally exciting and comes with more opportunities. This is also a good test when evaluating a prospective employer: do they have interesting data science problems that are central to their business? If not, you could find yourself far from the action.

AR: Q9. What key qualities do you look for when interviewing for Data Science related positions on your team?

ambiguityBK: Two very important qualities are a maturity in framing problems and building models and a comfort with ambiguity. The first usually comes with experience working with data and models. We look for people who can take a business problem and identify approaches for solving it with data. The best candidates are often the ones who know when a simple solution is the right one and don’t overcomplicate their approaches. Comfort with ambiguity is similarly important. Many people can excel when a clear problem is presented to them, but the best are those who can find the problem in the first place.

AR: Q10. What was the last book that you read and liked? What do you like to do when you are not working? the-everything-store

BK: The last book I read was “The Everything Store” by Brad Stone - a fascinating telling of the rise of Amazon. In my free time I enjoy hiking with my wife and dogs.