- BA, physics and computer science, Carleton College
- MS, electrical engineering and computer science, University of Minnesota
I am a distinguished engineer in AI working on innovative database architectures including document and graph databases. I have a strong background in semantics, ontologies, NLP and search. I am a hands-on architect and like to build my own pilot applications using new technologies. I also enjoy doing presentations and mentoring staff. I started the NoSQL Now! Conference (now called the Database Now! Conferences). I co-authored the book Making Sense of NoSQL, one of the highest rated books on Amazon on the topic of NoSQL. I worked at Bell Labs as a VLSI circuit designer where I worked with Brian Kernighan (of K&R C). I also worked with Steve Jobs at NeXT Computer.
- Graph technologies
- Deep learning
- Natural language processing
- Internet of things
- Image processing/computer vision
- Machine learning
- Intelligent agents
When I worked for Bell Labs in the 1980s, I received two awards for exceptional contribution to the area of automated VLSI layout using UNIX and C technologies.
- Keynote presentation at the 2019 Graphorum Conference in Chicago
- Keynote presentation at MedFuse 2019 in Minneapolis
- Presentation at GraphConnect 2018 in New York City
- McCreary D, Kelly A. Making Sense of NoSQL. Shelter Island, NY: Manning Publications, 2014.
- Ketter W, Batchu A, Berosik G, McCreary D. A semantic web architecture for advocate agents to determine preferences and facilitate decision making. Proceedings of the 10th International Conference on Electronic Commerce. Innsbruck, Austria, August 19–22, 2008.
Why did you choose to work at Optum (instead of any other tech or health care company)?
I chose to work at Optum for primarily two reasons:
First, for many years I would tell people I didn’t want to become a manager. I had owned a company with almost 100 people and I enjoyed building things more than managing budgets and making hiring decisions. When the distinguished engineer position became available at Optum, I realized I could keep advancing in my career and not have to manage budgets and other managerial tasks. This role allows me the freedom to work on technologies that have broad strategic impact for multiple business units and concentrate my focus on understanding the long-term impact of emerging technologies.
Second, I have the ability to access the right data sets, billions of records. The only way to quickly do analysis of billions of records is to distribute the workloads evenly over large numbers of computers. This is called distributed computing. Many developers at smaller firms never get to design distributed systems. At Optum I get the chance to design new distributed systems every month.
I am grateful for the work Optum has done moving toward a culture that allows technical staff to contribute to our product strategy.
Is there a personal story behind your decision to work in health care?
I was originally hired by MarkLogic because I had developed a methodology for matching business problems to the appropriate NoSQL database technology. When I was placed at Optum, I got exposure to some of the most complex organization and technical problems I had ever seen. I found a rich network of people who had a fantastic vision of lowering the cost of health care. I am grateful that my work in graph evangelism has had such a broad impact.
What’s one example of a really difficult problem in health care that you are solving using technology?
Getting all health care data in one large graph database and making real-time clinical decisions that leverage health care records for 50 million lives. By real time, I mean being able to crunch through billions of care plans in fractions of a second to find the right care plan for a patient. That means I need to compare a patient to 50 million other patients to find the most similar patients and recommend the next best action for a physician when he or she needs it most. We don’t want to wait for an overnight batch job to run. So we need to be extremely thoughtful about how we can divide the work over thousands of computers in a cost-effective way.
How are you keeping health care human while applying advanced technology?
Sometimes AI is used to make our advocates think like they know everything, even though they may have gaps in the vast health care landscape. We can do this by allowing our advocates to use a chatbot that has access to large knowledge bases.
What are you passionate about when you’re not at work?
Outdoors! Camping, biking, Ultimate Frisbee.
I am also a big fan of STEM mentoring and I am involved in the CoderDojo program as well as the co-founder of the AI Racing League that teaches AI concepts to high-school teachers and students. I have a strong interest in helping students from under-served communities get involved in technology. I am especially interested in getting girls and minorities exposed to fun and challenging problems through mentorship programs.
What advice would you give to your younger self?
Be patient, pursue your passions, and you will find your place in the world. When I was young I struggled with undiagnosed dyslexia. I didn’t feel as smart as the other kids. It took me a long time to figure out that my brain was not good at language processing but was ideally suited to visualizing data moving through complex databases and networks. I seem to be able to quickly visualize complex business solutions in my mind better than people with far superior skills than I have. Our brains are all different, and it just takes time to figure out what we are good at.
When it comes to solving hard problems, it is the collective power of your friends of friends that help you overcome obstacles. You don’t have to be the best at math or programming. You have to have the best network of people who will help you out when you need it.