I first got to know Amy Abernethy when I was an observer on the board of Flatiron Health, and she was the company's chief medical officer. Both of us were physician-scientists with a special interest in the data sciences, and we instantly developed a deep friendship.

One of the things that I especially admire about Amy is her continued ability to reinvent herself over and over again. She started out as a clinician, earned her PhD in informatics, and then went on to become a trialist. She was quick to realize that much of the future of medicine will be about data, and her leap from academia to Flatiron Health was all about improving the way real-world data could be used to accelerate medical research. And then she made yet another leap by joining the FDA, where she served for two years as Principal Deputy Commissioner.

In my own career, too, I seem to seek reinvention approximately every six years. I was obsessed with pure mathematics until I was 24. Then I applied to med school, and for the next six years I focused on basic biology. I spent the following six years on clinical medicine, and since then have devoted my time to building a software organization. I feel like things are coming full circle now, and in my new role at the Schmidt Center I'm returning to my roots in machine learning and mathematics, and their applications to medicine.

Physicians make very big decisions in people's lives. Whether or not to put in a heart valve, whether to do a stent or surgery, what chemotherapy to dose. We often make these big decisions with very little data for decision support. This has to change. As we discuss in this episode, the future of medicine will involve advanced machine learning to improve clinical decision making, and better ways to match patients to personalized therapies.