Brooke Cowan, senior director of data science for Apex Clearing, has studied spoken languages including Spanish, Japanese, French, German, and many more computer languages. Brooke earned her Ph.D. from MIT in computer science, studying machine learning and artificial intelligence. Before joining PEAK6, Brooke spent seven years at Expedia focused on how machine learning can support planning the perfect vacay. Now she’s putting her mad skills to work at Apex Clearing, making financial transactions–and financial markets–safer and more efficient. Here’s what Brooke had to say about machine learning, getting more women into computer science and more.
Right now:
As senior director of data science at Apex Clearing, my role is to identify use cases for artificial intelligence and machine learning, develop the vision, and hire a team to carry it out.
Why tech:
I actually was an undergrad American studies major and a teacher before going into data science. My dad inspired me to get into tech. As a medical researcher with an engineering degree, he encouraged me to build a website for his group at UCSF Hospital. I built the site and found that I liked coding. In some ways it was like learning a human language, to which I had always naturally gravitated.
Proud tech moment:
At Apex, I’ve been looking at how we can improve our processes around ACH withdrawal requests. This process is governed by a set of “auto-accept” or “auto-reject” rules. Any request that isn’t auto-accepted or rejected passes to a human analyst. My team has been training a machine-learning model on top of our analysts’ data so that we can reduce the volume of requests going through to humans. Our model can now make that decision with a relatively high degree of accuracy. I’m super excited about how this could improve our processes and enable us to offer new products and services to our customers.
Impossible made possible:
When I first joined Expedia, my team tried to build a natural language search engine for travel, rather than the form-based searches Expedia typically used. For example, we wanted customers to be able to say, “I want to go somewhere I can swim with dolphins,” or, “I want to go skiing in summer,” which is hard to express when you’re constrained to filters and check-boxes. Before I joined, my team had been trying to solve that problem with a rule-based approach, by parsing inputs. The first thing I did was build a machine-learning model that could identify the salient pieces of information, trained on examples. This was in 2012, and machine learning was new and exciting outside of a research context, and real-life applications were just beginning to ramp up.
When I’m not working:
I like to run and play with my kids. I also enjoy cooking and baking. Last night I baked naan.
Tech for good:
I’m passionate about getting women, particularly young girls, interested in science and math, and am excited about careers in these areas. Computer science graduates at the bachelor’s level are around 20% women. We can do better. Right now, I’m involved in a few PEAK6 initiatives supporting women in technology.
Best part of my job:
The people here are so nice, supportive and smart. Plus, I love the excitement around and investment in technology shared by the leaders of PEAK6. I feel lucky to be in a place where there’s support for the company’s values, not just in words but in actions.