On Friday, Jennifer Hill, who has a PhD in math and works for Capitol One, gave a seminar to our undergraduate students about working in industry with a degree in mathematical sciences (operations research, statistics, and math). Two decades of experience led her to give the students this simple advice: learn statistics and learn how to program. (I saw a lot of head-nodding from the faculty, so I don’t think the advice changes for an academic career). She also stressed the importance of operations research, since all of the problems she has worked on in industry involved limited resources.
Jennifer Hill’s advice to students was based on her experience in the home furnishing textiles industry, the chemistry industry, and the financial industry. She talked about a few projects she worked on to highlight what tools students might use for their jobs.
The first project, in the textile industry, was to evaluate three bids to build a manufacturing system. Each bid varied considerably, and consisted of very different designs and very different prices. She used simulation to analyze output statistics and to compare the designs before the company chose which bid to accept.
Another project, in the credit card industry, was to minimize the cost of fraud, including the cost spent to fight fraud. The objective was not to minimize the probability of fraud occurring, since Capitol One’s bottom line is cost. She also added that other criteria are considered, including the false positive probability (no one lines having a credit card charge denied—it’s unpleasant and embarrassing). The model predicts the probability of fraud at each point of interaction and is ultimately a game, since fraud is dynamic. And despite what the credit card commercials tell us, fraud is a rare event.