Full Width Image Background

Thomas Sullivan

Innovation Lead

Location: Eden Prairie, MN



BA, economics and philosophy, University of St. Thomas

Career bio

Upon graduating college, I took a position as a call center agent, helping attorneys and law librarians connect display stations to our mainframe computers to assist in legal research. I thought I wouldn’t last a year. Fast forward six years and I was managing 150 agents and losing a lot of sleep. Around this time, the commercialization of the Internet erupted and I was recruited to a small local company that was building a variety of call center software and poised to shake up the market.

I jumped and spent the next 18 years of my career with the firm as we continuously reinvented ourselves as opportunities arose and disappeared. During that I period, I worked in sales and marketing. I got the opportunity to launch our business in the U.K. where my wife and five children settled for two-and-a-half years and loved every minute of it. The kids learned how to play soccer and talk funny, I figured out how to drive a stick with my left hand.

Returning to the U.S., I added program management and product management to my resume before eventually joining Optum.

Personal profile

Why did you choose to work at Optum (instead of any other tech company or health care company)?

We’ve done some interesting early work around linking voice signals and clinical conditions. There is a strong body of medical and academic research demonstrating linkages between human speech production — the manner in which a person utters their language — and a number of disease types, including neurodegenerative conditions, cognitive impairment and mental health. An early and accurate diagnosis of these conditions could allow afflicted individuals and their families to begin treatments and interventions that may slow symptoms, reduce hospitalizations, control comorbidities and improve the patients' quality of life — all while helping reduce lifetime costs.

We’re investigating screening for the early detection of disease types by capturing their vocal biomarkers, and correlating their speech features to known disorders. Applications like these could measure the thousands of linguistic and paralinguistic vocal traits present in each conversation and submit this data, along with EMR and claims information, to a machine learning environment that can assess the likelihood the speaker has one or more conditions.

What are you passionate about when you’re not at work? 

I just finished the MS150 this weekend (a 150-mile bike ride to support the National Multiple Sclerosis Society). Biking that distance with my wife and one of my daughters was pretty cool. I enjoy traveling. I also brew beer. And bet on horses.

What advice would you give to your younger self?

I’ve already tried this. Flux capacitor completely failed. Had the darn thing worked, I’d probably given myself a lot of advice, like what stocks to purchase (Apple, UHG) and which ones to short (Kodak). Also bet the farm on the Minnesota Twins to win the World Series in 1987 and 1991. And move everything into credit default swaps in 2007.


Get the latest tech insights and deep-dives

Color Block



Join us

Help us solve the most wicked problems in health care through research, development and emerging technologies.

See open positions