Capital One Innovation, Technology & Agility

Describe a typical day with Capital One. What sorts of problems are you working on? What tools or methodologies do you employ to do your job?

I’m a ML engineer for the enterprise platforms and technology team. Our primary focus is solving complex ML problems that help our colleagues, who in turn assist Capital One customers in becoming more financially empowered.

We help teams solve problems using models we create from Capital One’s existing store of datasets. Our work can range from developing predictive time series models to implementing natural language processing. For me, most of this work happens using Python for data processing, internal platforms for orchestration of our applications and agile methodologies to help structure all of our work.

Share a project you’ve worked on that you’re particularly proud of. 

One of the projects I’m proud of is helping to build a suite of anomaly detection and monitoring tools that help us identify suspicious or anomalous activity involving financial transactions.

Product managers, data scientists and ML engineers came together to build models that better allow us to service our customers, such as more quickly mitigating fraudulent activity.

How is your team integrating AI and ML into the product development process, and what specific improvements have you seen as a result?

My team is responsible for investing in emerging research in the AI and ML space and applying it to the financial context. We do this by working closely with our tech partners and teams across Capital One to identify common modeling challenges within the business. Additionally, we accomplish this through applied experimentation to identify the best AI and ML techniques to solve a particular business challenge and scale the impact so it can be used across the company.

Our team has helped introduce unsupervised and graph ML techniques to support our customers and the business in various areas, such as identifying fraud, customer service needs and providing personalized app experiences. This work has helped reduce the time data scientists spend on customized feature engineering, allowing them to move quickly and efficiently deliver value to customers sooner.

What strategies are you employing to ensure that your systems and processes keep up with the rapid advancements in AI and ML?

My team works closely with our partners to deliver on Capital One’s AI and ML strategy, which includes thinking about responsible applications of AI. Across all of our work, we’re guided by a mission to build and deploy AI and ML in a well-managed way that puts people first. This includes working cross-functionally across the business to follow best practices, such as extensive testing and implementing human-centered guardrails before introducing AI systems or models into any customer or business setting. 

We also ensure that we only use AI and ML to solve a problem when it’s the right way to solve the problem. We see that, in some situations, business logic is sufficient enough to solve the problem at hand. We focus our efforts toward making an impact on our business rather than trying to use something flashy just for the sake of saying we use AI.

Can you share some examples of how AI and ML has directly contributed to enhancing your product line or accelerating time to market?

By partnering with teams focused on fraud, we’ve been able to use ML techniques to make big breakthroughs in our ability to mitigate fraud and help protect our customers. Using AI and ML, we’re better able to adapt to emerging changes in behavior patterns as fraudsters constantly change their techniques, greatly improving our detection of anomalous behavior. We’re continuing to experiment with new AI and ML capabilities to stay at the leading edge of this space and continue to give customers the best possible experience.