Analyzing Impact

Analyzing ImpactAnalyzing Impact

Rayid Ghani

Did you vote in the last presidential election? Rayid Ghani (ML'01) may have lent a hand in motivating you.

Creating algorithms to interpret mountains of gathered data, the Carnegie Mellon University alumnus and machine learning expert played a groundbreaking role as chief scientist for the Obama for America 2012 campaign.

Using analytics and machine learning, Ghani focused on using online tools like email and social networking to motivate people's offline actions, from fundraising to volunteering to voting. It was a daunting task made exponentially more so when scaled to the national level.

"Our goal was to make all the decisions of the campaign as evidence and data driven as possible," explained Ghani. "To measure the things we were doing and then allocate resources not based on the way people felt, but on evidence based on data we'd collected."

Ghani became intrigued with machine learning as a college junior in a small, liberal arts school. Looking to delve more deeply into the field, he spent a summer researching with Tom Mitchell, the E. Fredkin University Professor and chair of CMU's machine learning department. Hooked, he was back the next summer and stayed to earn his master's degree.

"Programming was telling the computer what it needs to do," he explained. "Machine learning was exciting because we were giving the computer the rules of the game and it would figure everything else out by itself."

Ten years at Accenture Technology Labs followed, working as director of analytics research with a variety of organizations and industries in a position that spanned academia and business. When Ghani was considering a move into the non-profit world, the Obama campaign came calling.

He was primed for the challenge.

"Campaign personnel said, 'We're doing some really interesting things with data and analytics, are you interested?'" Ghani noted. "So I asked myself, 'Was there anything else I could do in that 18-month period that would have more potential impact?' and it was a very easy decision."

Now that it's done, Ghani is picking up where he left off: committing to the non-profit world.

"Non-profits have a very small base of dedicated support and it's very hard to grow that base," he said. "My colleagues and I are expanding the work we did for the campaign in persuading supporters to get their friends and peers involved to see if we can do that for other non-profit organizations."

Ghani credits much to his CMU background, from his education to the "open and collaborative culture" to the valuable CMU network "that really helps you once you have left."

Throughout the years, he's remained connected. While at Accenture, he both funded CMU graduate students and collaborated on research with CMU faculty.

"One thing you realize if you're in this machine learning world is that everybody in that world is connected to somebody at CMU," laughs Ghani. "Anybody you meet, either they have spent time there themselves or have spent time with somebody who has."

Rayid Ghani visited the Pittsburgh campus on June 21, 2013 as the keynote speaker for the Big Data Analytics Workshop, part of the 2013 ACM Sigmetrics Conference held at CMU June 17–21, 2013.

Related Links: Machine Learning Dept. | School of Computer Science | article | article | Obama Launches Manufacturing Partnership from CMU