Dietrich assembled an ambitious team of researchers and oriented them toward some of the hardest problems known to applied mathematics at the time. Together, they developed a series of data-centric solutions to improve and even automate complex business processes — including resource allocation, workflow management and employee scheduling — for IBM and its clients.
In the late 1990s, the team helped Southwest Airlines rethink how it allocated flight crews. Until that point, a skilled employee had generated the airline’s flight crew assignments, but the work had become unmanageable. It wasn’t just a matter of navigating employee timetables. The work required adherence to union rules, pilot rest regulations and a general understanding of how much time it takes for a crew to recover between various flights. With so much to consider, the process of compiling schedules would often require more than a month.
Utilizing an emerging field of research, now known as machine learning, the team built an algorithm to solve the largest pairing problem in history. It provided the basis for a reliable scheduling system that saved Southwest millions of dollars. For Dietrich, it was the near-perfect challenge. “I want to be working on important, hard problems,” she explained to Think Research in 2003. “Problems that have the potential to change the way business is done.”
In 2007, Dietrich was appointed an IBM Fellow and in 2008 she became a vice president in IBM Research. She eventually became chief technology officer and strategist for Business Analytics and head of Emerging Technologies in the IBM Watson group, where she and her team continued to make inroads in optimization, simulation, machine learning and the instrumentation of new, high-value data.