When Professor Kyle Gallivan finally landed in FSU's math department in 2008, he felt right away that he had ended up in the right place. After more than a decade at Florida State working in computer science, he had a choice to make, a luxury he's grateful for as he looks back. "I took a look and said, 'maybe the math department is the right place,' and I was very, very lucky that the department agreed. I'm very happy that happened."
Gallivan seems to have a knack for winding up in the right place at the right time. His professional journey has taken him in a number of different directions, but in nearly every case, those turns have landed him at the leading edge of his field.
Gallivan grew up in the Chicago area, not far from some of the places where the beginnings of what is now known as computer science were just starting to take shape. That geographical coincidence may have ended up having a big impact on his career, because he doesn't recall having a natural predilection toward computers as a child. Of course, in the seventies, opportunities for young people to get hooked on computers were rare. "My high school had the old teletype connection to the university I ended up going to," he recalls. "I remember doing Fortran and BASIC programming on that thing when we could... but I wasn't particularly interested in mathematics as it related to computer science."
That teletype connection was to Lewis University, a small college just outside Chicago, where Gallivan went for his undergraduate degree. Lewis was located near Argonne National Laboratory, and Gallivan found himself working there as a student research associate in the Reactor Analysis and Safety Division. That placement was arbitrary, but the work that came with it caught Gallivan's interest. "I didn't do any of the hardware," he explains, "I helped implement and then run and maintain the numerical simulation... obviously I wasn't responsible for the modeling, but it was my first exposure to this nice mix of experiment, simulation, and significant engineering analysis."
"I took a look and said, 'maybe the math department is the right place,' and I was very, very lucky that the department agreed. I'm very happy that happened."Kyle Gallivan
His work had to do, specifically, with running simulations to learn more about the effects of changes in coolant conditions with nuclear reactors. It was important work, but not something that was getting a great deal of attention at the time. Not long after, however, the Three Mile Island disaster occurred. Gallivan himself was still an undergrad, and his time at Argonne was coming to an end, but he remembers the excitement of being involved with research that ends up being of vital, practical importance. "All of a sudden, everybody was talking about loss-of-coolant accidents." Many of his coworkers, in fact, went on to work on reactors. Gallivan decided to go to grad school in computer science.
He wound up at the University of Illinois at Urbana-Champaign and found teachers that would help shape the rest of his career. "There was a magnificent group of professors at that time there, pioneers of computer science across the board - hardware, architecture, software." Among those faculty members, he found in Professor Bill Gear a mentor who taught him the importance of understanding the bigger picture. "He reinforced even further the need to do theory, to produce reliable software based on your algorithms, and use the theory and the software together to understand how things work."
After graduating with his Ph.D. in 1983 Gallivan did a short stint in the private sector, but found his way back to UIUC within a couple of years. He was drawn back by his friend Ahmed Sameh, who had been on his Ph.D. committee, to work at the Center for Supercomputing Research and Development (CSRD). Researchers there were working on what was called the CEDAR System, an experimental supercomputer. "It was a unique place in the history of computer science and computation. We actually did true multidisciplinary work," he recalls. It was exactly the sort of work Bill Gear had emphasized when Gallivan was a graduate student. There were people at CSRD working on all aspects of the project, from the theory and planning to the building and testing, and while people specialized in one area or another, everyone participated to some degree in every phase.
That kind of integrated cooperation is a rarity these days, when different aspects of computing research are done by independent groups, and Gallivan misses the advantages of working with people who have a broad understanding of what they're working on. "It was beautiful, absolutely beautiful. I don't think such a place has existed since then." He gives a lot of the credit to Professor Sameh, who was his supervisor at CSRD and is still a close friend, for the unique environment.
Argonne National Laboratory in Argonne, Illinois, where Gallivan first participated in large-scale interisciplinary work on computer systems.
Perhaps the greatest testament to the quality of the work he was involved in at CSRD is the fact that many of the modern technological devices that have revolutionized the world can be traced back to projects developed at UIUC and other sites doing similar work 20 to 30 years ago. The advances they made ranged from large, high performance systems down to highly efficient devices for digital signal processing and communications. "Those are the precursors of cell phones," he says. "I mean, the digital revolution occurred at that time."
CSRD eventually closed as the university changed its focus, and Gallivan moved into the Department of Electrical and Computer Engineering. At that time, the earliest university computational science programs were beginning to form in scattered places around the country. One of those universities was Florida State, where Professor M. Y. Hussaini had founded the university's Computational Science and Engineering program (the precursor of today's Department of Scientific Computing). In 1997, Gallivan took an appointment at FSU, largely because of Hussaini's involvement.
He brought with him work he'd begun while at Illinois on model reduction. Model reduction involves finding ways to write new, smaller systems of differential equations to approximate larger ones, thus reducing the time to solve them. Over the years he managed, collaborating with Professor Paul Van Dooren and graduate students, to develop some techniques for reducing certain differential equations that are now widely used in the field.
Over 13 years at FSU, the focus of his work has shifted some, but, whatever the project, he's always found just the collaborators he's needed. Over his career as a teacher at FSU and UIUC, he's found several exceptionally bright master's and doctoral students who have been invaluable to his work. "You tend to learn more from the good students than you teach them, ultimately." Several of his students, including undergraduates, have gone on to earn doctoral degrees and start making their own contributions to the field.
By 2008, when Gallivan made his move from computer science to the Department of Mathematics, he had already accumulated plenty of experience with the culture of FSU and found it a great place to continue his work. That move, however, turned out to be yet another good one, and he's found among his new colleagues a supportive academic community. "It's a very welcoming department. I was very happy that they said 'sure, come on.'" In exchange for that supportive community, he fills an important niche for the department, teaching courses in the foundations of computational math and other subjects as needed. Teaching is not only work, but something he finds enjoyment in.
Gallivan and his family - he's married and has two daughters now in their twenties - have also found in Tallahassee a pleasant place to call home, despite the downsides of being transplanted from Illinois. He was once an avid hockey player, but the Florida culture has put a damper on his athletic pursuits. "They used to have a minor league hockey team, so they would freeze the civic center. That's gone by the wayside, and I'm not going to skate on synthetic ice. That shouldn't be allowed."
Fortunately, between teaching the future minds of computational math and paving the way for new advances himself, he's got plenty to keep him occupied.