My professional and educational background is actually in computer science – I transferred from RGA’s IT department to the data analytics group in 2013. So I came into the data side of things by way of technology. I love math, and I love to solve puzzles. The math involved in data science is complex, challenging, and constantly evolving.
Jeff Heaton, Ph.D., FLMI, is Lead Data Scientist for RGA Reinsurance Company’s Global Research and Data Analytics department. He primarily focuses on modeling risk for underwriting systems using electronic health records. He also works to bridge the gap between complex data science problems and proven software. He is a frequent speaker and author for organizations such as the Society of Actuaries (SOA) and the Institute of Electrical and Electronics Engineers (IEEE), and for academic journals. Jeff has authored several books on artificial intelligence, and he teaches a graduate course at Washington University in St. Louis on this subject.
We sat down with Jeff to discuss his career, artificial intelligence, and the future of Big Data.
My professional and educational background is actually in computer science – I transferred from RGA’s IT department to the data analytics group in 2013. So I came into the data side of things by way of technology. I love math, and I love to solve puzzles. The math involved in data science is complex, challenging, and constantly evolving. It has been a good transition: Coming from an IT background allows me to ensure the data science team effectively works with RGA’s IT department. An understanding of computer science helps me to leverage the amount of data we have available today. This offers an unlimited amount of problems to solve, and unlocking the insights behind the data fuels my passion for the job.
In 2001, RGA was launching some innovative projects on the internet. The internet was not so common in business back then, but was often discussed at technology clubs and organizations. I learned about the projects through RGA’s Chief Information Officer at the time, who belonged to some of the same tech clubs and organizations that I did. Though I started RGA as a web developer, I quickly found my passion for implementing the complex algorithms that run RGA’s retrocession and premium calculation engines. I had always been interested in artificial intelligence (AI), and published on it. When RGA needed AI programmers, I was able to leverage my AI knowledge to join the data science group. So I essentially turned a hobby into a career. I’ve been at RGA ever since.
Insurance combines so many different types of data – financial, medical, behavioral. The sheer volume of information is simply too much to analyze effectively without some form of cognitive computing power; it requires systems that can make thousands of calculations per minute. When applied to underwriting, for example, AI can very rapidly triage business by accepting or rejecting certain applicants based on a given set of criteria. The future of AI and underwriting will take this many steps further, applying new data sources and scoring models to better refine accelerated decisions so that a greater number cases can be processed automatically. The industry’s use of credit data as underwriting evidence offers a powerful example. Developing robust, secure AI systems to process this information is essential. It is a case of “Big Data” meeting “Big Compute.”
I’ve heard data referred to as “the new oil” in that companies can make money through data in ways we’ve never thought possible. But money cannot be the only motivator; we must harness the power of data to develop solutions that benefit our customers and help move society forward. Profit and progress need not be mutually exclusive. To thrive in the data-driven economy, companies must deliver profitable solutions that also make people’s lives better. For example, insurance-linked wellness programs offer lower premiums for better fitness by using wearable technology that gathers biometric data. Regardless of the application, personal data must always be used responsibly; the privacy and wellbeing of each individual must be the top priority.
Well, you better like math. And you better be flexible – willing to abandon your original solution for any better one that comes along, no matter how much work you’ve already put in. There are no sacred cows in this field, and you can always improve your model. The key is to make sure the conclusions you draw from any set of data make sense, which requires working with subject matter experts. All that said, it’s a wonderful profession for the right person. I get to work with the latest technologies and with an extremely intelligent and talented team of colleagues pursuing a common mission. Best of all, no day is the same. There are no one-size-fits-all, cookie-cutter solutions, so you get to blaze your own trail on any given project. I love what I do and know that I am making a difference. Isn’t that what everyone wants from a job?
