From Launching Satellites to Banking Security Innovation
How Dave Castillo’s Passion for AI/ML Led Him to Capital One
David Castillo, Ph.D. joined Capital One this summer as our new Managing Vice President of Machine Learning. In that role, Dave will lead Capital One’s research and strategic innovation initiatives in artificial intelligence and machine learning (AI/ML), and look at how AI/ML can be used in different areas of the business to make banking more human, intuitive and enjoyable for customers. He’ll also work on the development of tools, technologies, frameworks, and partnerships with industry and academia.
We sat down with Dave to discuss his background, his passion for AI/ML, and his decision to join Capital One. The following are excerpts from our conversation.
You are an expert in artificial intelligence and machine learning and yet you chose to work in banking and specifically at Capital One. Why is that?
For me, it was a no-brainer. Banking is the next frontier for AI/ML, and Capital One is known for being a leading innovator — not just in the financial services space, but in the broader tech sector. They’re using AI/ML for everything from identifying security threats in a constantly evolving banking environment, to improving the customer experience at call centers. I knew this would be a great opportunity to create incredible products that make customers’ financial lives easier and more enjoyable, and so far, it’s been a blast.
Personally, coming to Capital One was a very natural progression. After stints earlier in my career at NASA and launching satellites with Motorola, I joined Early Warning Services in 2015 and worked with financial institutions on technology-driven payment and risk solutions. I was the head of their data science, data management platform, and innovative labs group. This included working with a consortium of some of the largest U.S. banks, Capital One included, and we helped build Zelle, the peer-to peer payments platform.
I worked closely with Capital One at Early Warning Services, and everything the team did — from their personalities, to the way they thought about data innovation — was so far advanced that they really left an impression on me. I had always known that Capital One was innovating at a pace that rivaled tech companies like Airbnb and Netflix, and the team had visionary technology leaders like Diane Lye [Senior Vice President in Enterprise Data Services] who helped create Capital One’s data ecosystem — I wanted to be a part of what the team was doing, much of which hasn’t been done in the industry before.
Tell us more about working at NASA and your experience with satellites.
After completing my undergraduate degree in engineering, I decided to pursue a Master’s degree in decision sciences with a software focus. From there I was exploring Ph.D. programs and ended up receiving an offer to work at NASA.
I initially deferred, wanting to focus on pursuing a Ph.D., but after they invited me to their Kennedy Space Center in Florida, I became totally seduced by the possibility to work with applied artificial intelligence.
As a young engineer with NASA, I was using a lot of different systems, and earned my Ph.D. at the University of Central Florida. It was all fascinating work, but I became less fixated on academia as a career path and more interested in the application, deployment, and wide-scale adoption of the projects I was working on.
Working at NASA also forced me to get up to speed on distributed computing.
Around this time, Motorola had announced they were launching a network of over 60 satellites for a massive celestial and terrestrial project that would leverage streaming data. They were looking for a chief software engineer to lead the work. I applied, and was astonished to get the job! It was another pivotal opportunity that pulled together my experience with NASA, distributed computing, and streaming data.
What personally excites you the most about the world of AI and machine learning?
I have spent a great deal of my professional life as a practitioner of AI and machine learning, meaning I like to work with data, build solutions, and solve problems using AI/ML. Using this experience and passion for innovation, I am looking forward to enabling (and being a member of) a team that delivers value-added AI/ML solutions across the organization. One such enabler is an area that I am personally very passionate about and is what I like to call driverless data science.
Driverless data science is ultimately about self-learning models, beginning with model formulation through model delivery, which enables greater efficiency and speed throughout the machine learning process. Even while we make great progress towards this, we have human-centered processes and controls throughout every step of our machine learning environment, and I’m elated that I can help contribute to advancing this at Capital One.
What will you focus on in your new role at Capital One?
I’ll be looking for sophisticated ways to continue to leverage AI/ML and data engineering in our business processes while creating best-in-class customer experiences, going beyond where we are even today.
I’ll also look at how we can constantly strive to optimize our engineering and modeling processes to ensure that teams across the business can harness them for any range of use cases.
You once considered a career in academia and still work as an adjunct professor. How will you pair that interest in academia with the work you are doing at Capital One?
It’s really important to stay on the cutting edge of AI, machine learning, and data science. We’re looking to partner with top universities on compelling, high-value problems that will propel the industry forward.
One example is the work we’re doing in Explainable and Fair AI. We’re working with various university partners so we can maintain the highest standards for explainability — in an ethical and fair way — as we develop more advanced models for more use cases. We think that our efforts in researching explainability and fairness can help inform and contribute to how machine learning is developed and implemented across the industry.
Working as an adjunct professor keeps me current in terms of my ability to code, learn new languages, and stay on top of open-source developments — you really need to be on your game to teach students! I try to introduce a breadth of practical, applied use cases and projects so it’s not all theory, as well. Teaching is a nice escape that feeds another side of my passion for AI/ML.
Combined, these efforts can only enhance the state of AI/ML at Capital One and across the industry as a whole. Of course, the end goal is to create intuitive technologies that give people the tools to make their financial lives simpler, more enjoyable and easier to manage. I’m thrilled to be a part of this innovative moment at Capital One.
DISCLOSURE STATEMENT: These opinions are those of the author. Unless noted otherwise in this post, Capital One is not affiliated with, nor is it endorsed by, any of the companies mentioned. All trademarks and other intellectual property used or displayed are the ownership of their respective owners. This article is © 2018 Capital One.