Introducing Keith Marzullo
Donna M. and Robert J. Manning Dean
Dr. Keith Marzullo has been named the Donna M. and Robert J. Manning Dean of the UMass Amherst Manning College of Information and Computer Sciences (CICS), effective July 1, 2025. He joins CICS from the University of Maryland, where he served for nine years as dean of the College of Information. There, he led strategic growth, expanded academic and research programs, and deepened partnerships across academia, industry, and the community.
Previously, Marzullo held senior roles in the federal government, including director of the Networking and Information Technology Research and Development Program at the White House Office of Science and Technology Policy, where he coordinated R&D across more than twenty federal agencies. He also served as director of the Division of Computer and Network Systems at the National Science Foundation. Earlier in his career, he held faculty appointments at UC San Diego, the University of Tromsø in Norway, and Cornell University.
An ACM Fellow and a leader of the international iSchools organization, Marzullo’s research has focused on cybersecurity, privacy, and socio-technical systems that support human well-being. He holds an AB in physics from Occidental College, and both an MS in applied physics and a PhD in electrical engineering from Stanford University.
In August 2025, the Significant Bits editorial team met with Dean Marzullo for a wide-ranging video interview. The excerpts below come from a lightly edited transcript of that conversation.
What was your dream growing up and how did that inform your career path?
My dream when I was growing up was to be an astronomer. I loved astronomy—when I was a teenager, my father and I built a telescope together. We ground the mirrors, built the mount, built a clock drive, and took it out to star parties. I don’t do a lot of astronomy anymore, and I certainly don’t have the time to grind mirrors. But it awakened in me an interest in understanding how things work: “What is the way the universe works?” “What are the rules that it follows?” And that pulled me into physics.
When I was a PhD student in solar astrophysics, there were no jobs. I transitioned into electrical engineering because I found the topic of computer systems interesting. Much like physical systems, computer systems have rules, and trying to understand how they work pulled me in that direction. I was especially interested in concurrency—how things running together work, much like stars and planets—and that moved me into distributed systems.
Later in my career, I realized that the most complex systems with real concurrency are those involving people. And so, my path from wanting to be a lonely astronomer to working with people was long, but it was always driven by trying to understand how things work.
You began in physics but found your way into computer science through early programming work at Occidental College and NASA’s Jet Propulsion Laboratory. How did that pivot lead to the research that became Marzullo’s algorithm?
My undergraduate degree was in physics at Occidental College, a small liberal arts school, but I supported myself through programming. I worked in the data center at Occidental as well as at NASA’s Jet Propulsion Laboratory, where I programmed software for the Viking Orbiter and the Voyager missions.
When I realized there were no careers in solar astrophysics, I switched to programming and computer science in the Department of Electrical Engineering at Stanford University. The person I wanted to work with, Susan Owicki, was a professor in electrical engineering and an expert in concurrency.
I did much of my PhD work at Xerox Palo Alto, where they had invented much of distributed computing. I was part of the systems development division. They were implementing operating systems, and what we needed was a clock-synchronization system.
I defined the problem as a PhD student—no one told me to do it, but it was a lot of fun. We had clocks that would work well, clocks that would stop, clocks that would run backward. Trying to build an algorithm that would keep them tightly synchronized in the face of failures was my challenge, and that’s what got me into fault-tolerant distributed systems.
I developed something that’s now called Marzullo’s algorithm. Technically, it’s triple modular redundancy voting done on continuous values. It’s a way to determine, when you have multiple physical values, which are correct and which are not.
At one point, it dawned on me that the algorithm would not only work for synchronizing clocks but could be used for any kind of physical process, including synchronizing the rates of clocks. In other words, I could take a set of clocks and make them run at the same value and the same tick rate using a different version of that algorithm.
How has the computer science field changed since your early research days, and what challenges or opportunities do you see defining the next era of computing?
The application of computer science has been growing since the early days, when it was primarily concerned with technologies that worked with other technologies. Computer science is now a developing technology that works with people—algorithms and techniques are becoming intimately involved in our lives. It involves people and society, and computer science is stepping up to address that.
The concept of socio-technical design has become a major theme in computer science, which asks: “How do you design technology that works with people and understand how the policies of people affect the way they use technology?” That kind of authentic co-design is now very important, and it’s something we do well here at CICS—it has become central to what we do.
What excites you most as you step into your role as dean of CICS?
I'm excited about being the new dean of CICS because of all the great work that’s going on here. There’s an incredible collection of faculty, and the motto of the college—Computing for the Common Good—really resonates with me. Much of the work we did in the federal government was towards that aim. The things I was doing in my previous job were towards that aim. And this college is really serious about that. I'm excited to help this college grow even more in that direction.
How do you approach the idea of “Computing for the Common Good”?
If you're computing for the common good, how do you know it's for the common good? That can’t be just a gut feeling. You have to be guided by different disciplines and ways of understanding the world.
Technology is one part—it’s a value-independent tool that opens new possibilities, for good or harm. The social sciences offer insight into how people and communities respond to these technologies. The humanities help us examine the ethical and moral questions at the heart of what “the common good” really means.
This college has ties across all three areas. What I'm excited to do is fan that fire, grow it even more, and create more connections with the university, with the social sciences, and with the humanities, while continuing to invest in strong technology.
What do you see as CICS’s strengths when it comes to emerging technologies?
Computer science is always changing. The problems we're addressing are always changing. Artificial intelligence and machine learning are among our strongest research areas, and they’re transforming our society. But we're also strong in many other areas—networking, information systems, information retrieval, quantum computing, robotics, among others.
One of the things I love about UMass Amherst is that it’s a public R1 university. Massachusetts has many excellent private institutions, but UMass is the R1 public university for the Commonwealth. We have a reach toward a much broader set of people, helping them to be successful in computing for the common good.
The opening of the new Computer Science Laboratories building is a major milestone for the college. How do you envision the building shaping collaboration and community?
In this post-COVID period, the way people come together and work has somewhat disintegrated. Creating a new space where people can come together, be creative, work with students, work with the community, and engage in both education and research is
very exciting.
This new building is perfect for that. It has wonderful spaces for classrooms, for interaction, for students, and the new Physical Computing Makerspace offers incredible opportunities for hands-on learning. The building is going to create an incredible opportunity for our faculty, staff, and students to compute for the common good.
As you look ahead to your first year as dean, what are your top priorities for the college?
An idea without a plan is just a dream. Now is the time to ask ourselves: “Where are we going?” “What do we want to be over the next ten years?”
The things we need to teach our students have changed. The important research topics continue to evolve, and we're going to stay true to computing for the common good. But we need to understand where that’s going and how we're going to get there—in research, in partnerships across the university, and in working with the community, whether it’s the Pioneer Valley, the commonwealth, or the nation. It's going to be a lot of fun.
Planning is a great process because it allows you to dream. Then it allows you to determine a concrete way to get there.
This story originally appeared in the Winter 2025–2026 issue of Significant Bits magazine.