STEM Career Spotlight: Alyssa Frazee, Machine Learning Engineer

Jan 19, 2016 | Westwood

Every day, Machine Learning Engineer Alyssa Frazee uses math to fight crime... and we have to admit, her job title sounds pretty cool, too! In the second installment of our STEM Career Spotlight, learn how talented professionals like Alyssa create and use math models to detect fraud.


Alyssa Frazee

Occupation/Job Title:

Machine Learning Engineer at Stripe

Because job titles only say so much, tell us more about your job!

I'm glad you asked! I work at Stripe, a company that provides a way for online businesses to accept credit card payments. The machine learning team is responsible for building models and software to detect and block fraudulent payments, so basically, I'm crime fighting with math! :)

What do you love most about your job?

I love the variety of work I get to do. I might spend one day writing SQL queries to answer an important question our finance team has, and I might spend the next day writing Ruby code to help detect new payment patterns across Stripe. I also love my coworkers! People that do machine learning come from all kinds of different backgrounds, and it's interesting to hear a wide variety of perspectives.

How often do you use math in a typical work day?

I suppose it depends what you mean by "math," but I'd say the answer is "very often!" Our fraud models are based on a solid math foundation, so figuring out how to improve those models always involves at least a little bit of math. I also often work on figuring out how to measure stuff: how do we measure whether our models are doing well? How do we measure how much influence a single charge or business will have on our overall systems? How do we measure whether our model improvements are effective at foiling fraudsters? And how do we measure how sure we are about any of this? All of this involves a bunch of math (and statistics!) so it's a big part of my everyday routine.

Let's backtrack a little. Fill in the blank: When I was little, I wanted to be ___________ when I grew up!

In kindergarten I dreamed of being an artist (specifically a children's book illustrator). By middle school I'd changed my mind and planned on becoming an architect that designed fancy houses. And in my early 20s I'd planned to become a researcher, specifically using math to better understand tough medical research problems like cancer. 

Did you like math when you were younger?

I mostly did! I didn't always love my classes in school (they could get a bit tedious) but I thought a lot of concepts related to math were interesting. In elementary school I learned about a number called "a googol", which I thought was the most hilarious thing, it was really cool to see where the number pi came from, and in my teens, I read a really neat book about the history of math being used to send coded messages and break the codes of other peoples' secret messages. 

Which components of your education (formal or informal) helped you get to where you are today?

What has really helped me is having to solve a problem that nobody has ever solved before. For me that came in the form of getting a PhD (I studied biostatistics in grad school) and working on some new programming projects in my spare time. Until grad school, I was mostly working on homework problems that lots of people before me knew how to do, but when I started doing research after college I realized I was trying to answer questions that nobody had ever answered before, which was initially very scary because what if they were impossible questions?!

However, working on those kinds of problems helped me learn how to take very small steps toward a solution and helped me be okay with sometimes being completely confused for days. It helped me get better at asking good questions and brainstorming ideas with friends. It's also helped me not freak out if I'm stuck on a problem: I've been stuck a lot of times before, so I mostly trust that I can eventually get un-stuck. :) 

What do you find most fulfilling about your work?

I really like working on fraud because it's interesting from a math perspective (it's sometimes quite hard to tell if a payment is from a fraudster!) and it's very practical: most people are interested in doing some kind of business on the internet these days, so I know that the work I'm doing is impacting people right this very minute. I'm also learning forever! I went to four years of college and five years of grad school and I STILL learn new stuff at work every single day, which rocks.

What advice do you have for kids interested in pursuing a STEM-based career?

I'm so excited to see what you do! My advice is to keep at it, if you like it! STEM fields move quickly and there are always new types of jobs being created (my job didn't exist 5 years ago), so keep your eyes open for interesting things.