Math-to-Industry Boot Camp: two mathematicians’ journeys to mentorship

MINNEAPOLIS / ST. PAUL (07/28/2023) – In the summer of 2018, Christopher Miller was about to enter the final year of his PhD program at the University of California, Berkeley. He knew he wanted to pursue a career in finance, but needed to expand his technical and programming skill toolkit. Miller attended the Institute for Mathematics and its Applications (IMA) Math-to-Industry Boot Camp later that summer and discovered his passion for machine learning.

“The IMA was hugely impactful. Before attending, I had planned to work in finance, but the IMA helped me develop critical machine learning and programming skills that are useful in tech. After the program I realized how much I liked machine learning, and that I wanted to pursue a career in it. The summer after attending IMA I graduated and got my first tech job at Ancestry, which I attribute largely to the IMA experience,” he says when he reflects on his Boot Camp experience.

Today, Miller is an Applied Researcher at eBay. His team utilizes machine learning to extract information from customer listings to organize the website’s massive user-created inventory in intuitive ways. The output of the machine learning processes is passed to the eBay search engine to make it easier for the site’s shoppers to navigate the platform. When Professors Daniel Spirn (University of Minnesota) and Thomas Hoft (University of St. Thomas) reached out to Miller about mentorship earlier this year, he says he was happy to help.

“I wanted to give back to the program that had helped me so much. Also I wanted to see if the students could crack the particular problem I gave them, which I consider to be quite hard. They have done a great job so far. My favorite part of being a mentor has been seeing the creative and sometimes brilliant ideas proposed by the students,” Miller says. For the past few weeks he’s been guiding a team of six graduate students through the “Multimodal Search in eCommerce” capstone project. The group is exploring solutions in meshing image-based and text-based search features with a two-phase multimodal search. 

Dr. Leo Digiosia, quantitative model analyst for US Bank,  is one of the co-mentors on the “An Excess Demand Model of Home Price Appreciation” capstone project. 

“As a quantitative model analyst, my work around building economic and financial models is highly research-driven. As one does in a PhD program, our team answers important questions that require combinations of creativity and mathematically quantitative attention. For example, what historical relationships have we seen between mortgage rates and Treasury bond yields? How can we explain recent divergence from these trends? What types of forces in the housing market  contribute to rising or falling home prices? As an analyst working in the mortgage division of corporate treasury, these types of questions are interesting to me. Their answers inform investment portfolio decision making and interest rate risk hedging at the bank,” Digiosia says about his work. The biggest lesson he’s learned so far in his position? “Primarily, I’ve come to appreciate that the hardest thing I’ve done is getting a PhD in math.”

Digiosia, too, was a Math-to-Industry Boot Camp attendee just last year, working on a project led by US Bank, the same company that would hire him six months later.  “I was surprised to find in the six week program how easy it is to pick up coding languages like Python and R. While I do believe that the intense boot camp fast-tracked my coding ability, I think many math graduate students would be surprised at how naturally they can learn these skills. As I mentioned, we learn how to learn during the PhD program and this is an unparalleled skill to have.”

The team of five graduate students working on Digiosia’s capstone project are working to create a story around an excess demand model of home appreciation. By pursuing the topic through the lens of differential equations, the team can forecast values of home price appreciation on a variety of time scales. Digiosia says bringing this project to Boot Camp has been a great way to better establish the conceptual soundness of this model he has been developing for months. “To see the team become immersed and interested in the project and make great progress has been a delight,” he says.

For the past eight years, the IMA Math-to-Industry Boot Camp has served as an intense six-week session designed to provide graduate students with training and experience that is valuable for employment outside of academia. The program is targeted at PhD students in pure and applied mathematics, and consists of courses in the basics of programming, data analysis, and mathematical modeling. This year, there are five UMN Math PhD students participating in the Boot Camp. The program is organized by UMN Professor and IMA Director Daniel Spirn in collaboration with University of St. Thomas Professor Thomas Hoft. For students like Miller and Digiosia, the Boot Camp is more than just a summer experience; it’s an opportunity to give back to the mathematics community.

What advice would the two give to students and postdocs looking for/applying for post-UMN jobs?

Digiosia says: “Look for employers that have a clear appreciation for the work that goes into a PhD program. Even better, look for a hiring manager that has a PhD themselves – they’ll understand that there is little that a new hire with a doctorate cannot do, or cannot learn in a short amount of time.”

Miller says: “Lots: Learn the basics of data structures and algorithms (roughly 2 semesters of undergrad coursework). Learn to program in Python (you should be able to implement any data structure that you can envision). Take a graduate or advanced undergraduate ML/CS course. Do practice interviews with your friends. Do Leetcode problems. Seek internships. Apply to many jobs.”