VIEW THE RECORDED TALK

Jennifer Tour Chayes got to where she is by hard work and by maintaining a readiness to pounce when opportunity knocked.

“If a brass ring comes along, you grab it. You don’t question. You just grab it,” she said.

For Chayes, the biggest brass ring of all was an invitation two years ago to start a cutting-edge research lab for Microsoft Corp., her employer. She discussed this venture and her current role as its managing director in an April 20 address to HMC alumni, faculty and students gathered for the Annenberg Leadership and Management Speaker Series.

“I did not set out to be a leader,” the Princeton-, Harvard- and UCLA-trained expert in discrete mathematics and computer science insisted. “I set out to want to do science.”

She later added, “If you want something, it’s not enough to just have a vision. You have to tie it into other people’s vision.”

‘An age of networks’
Chayes’s vision was to launch a research center that would bring together the best and brightest from the fields of economics, math, physics, biology, medicine, sociology, anthropology and computer science to collaboratively develop new high-tech products and applications under the Microsoft banner. After winning the go-ahead to establish such a facility, she said she persuaded the company to site it in Cambridge, Mass., because that was where the topmost talent in those various specialities were to be found – and where they desired to remain planted.

But Cambridge also turned out to be an ideal home for Microsoft Research New England because so many potential clients were located nearby, particularly biotechnology enterprises. Said Chayes, “There is a revolution taking place in biology....If you know algorithms and if you know mathematics, you really can make a contribution in biology.”

For instance, a leading Boston-area cancer institute plans to utilize Chayes’s lab for computer-based testing of experimental oncology drugs and related therapies.

Chayes went on to tell the audience that a lab such as hers has special merit because this is “an age of networks.” Networks, she said, can range from the purely technologic to the economic, the biologic and the social. “There are also general resource-allocation networks,” Chayes added.

Working hand-in-hand
Owing to the existence of these varied and many networks, it is no longer possible, she contended, to “separate economic, social and psychological insights from mathematical models.” Neither can engineering be separated from business models. “Years ago it was that you did engineering, you came up with products and somebody figured out how to sell them,” she said. “Now there’s a different [approach] where you invest billions of dollars and then try to figure out what the business model is – also not a great way of doing it. [Instead] what you should be doing is working hand-in-hand, the engineers [with] the people who are coming up with the business models....”

Another current initiative of Microsoft Research New England involves the analysis of large-scale networks, notably those pertaining to resource allocation. Chayes reeled off some worrisome statistics: “[D]ata centers in the U.S. use 2% of the total power in the United States. That is doubling every five years. That cannot continue....[We have to] find out how to do it much more efficiently than we’re doing it now.”

Cloud computing could be an answer, she suggested, noting that future technology and business-model improvements surrounding this data-management strategy could eventually cause today’s in-house IT departments to go the way of the buggy whip. But first, it will be necessary to figure out how to incentivize the migration from in-house IT to cloud computing, Chayes offered. “We can’t separate the incentives from the goals, which I think has always been true when you want to incentivize various kinds of behavior,” she said.

Chayes closed by saying “Everywhere we look we see large-scale networks, [modeling and analysis of which are resulting in] new theories, theorems, experimental predictions, some new business models, possibly new drug therapies. And, of course, this is just the beginning.”