The life of Ümit V. Çatalyürek, the newest faculty member at the Georgia Institute of Technology’s School of Computational Science and Engineering and the new associate chair for academic programs for the school, is one in constant pursuit of novel and stimulating challenges.
Along with 20-plus years of teaching and research experience, Çatalyürek’s research interests lie in high-performance computing (HPC), also known as supercomputing.
However, if not for a fortuitous gift back in his homeland of Turkey, Çatalyürek might have been working with blueprints instead of computer chips.
“Early on, it was clear I was a math person and liked building things. So, I was planning to study architecture,” he said. “Then, I got my first computer in 10th grade and realized the possibilities of creating things were endless with a computer.”
After ranking among the top 200 students on Turkey’s National College Entrance Exam – Çatalyürek enrolled on a full scholarship at Bilkent University, home to Turkey’s top computer engineering program. While there, he earned his undergraduate, graduate, and doctoral degrees in computer engineering and information science.
Çatalyürek has not only explored his computer science interests but entrepreneurial ones as well, founding and managing a successful startup with his friends while at Bilkent. They developed the Turkish version of Microsoft Office. At the time, the software was one of the few on the market and accompanied every IBM computer sold in Turkey.
Although the business was successful, Çatalyürek longed for grander challenges than those posed by the business world.
“[The business] was fun, but I felt the industry would become repetitive for me ... and I really like challenging myself,” he said. “For me, the problems that held the most interest were in academia, not business.”
Although Çatalyürek’s curiosity for research was clear, he had yet to select an area on which to focus. However, that started to shift as he transitioned into his graduate studies.
Unlike most students, who chose a graduate adviser based on whether that professor’s research aligns with their interests, Çatalyürek selected the adviser with whom he thought he could work best – an adviser who happened to be studying HPC.
“I was always interested in getting the most out of a computer. Yet, I did not consider researching HPC until I selected my adviser,” he said. Working alongside his adviser, Çatalyürek’s passion for HPC and combinatorial scientific computing research areas grew.
He admits his foray into supercomputing was not an “informed decision.” Nevertheless, he credits it as a pivotal moment in his career.
“Throughout one’s life, we have many moments that shape our lives. The experiences that profoundly shaped my life were the years working with my adviser.”
Now with 20-plus years of teaching experience under his belt, Çatalyürek continues to emulate that same tight and supportive relationship with his students.
“My teaching philosophy is always to have fun,” he said. “I had a lot of fun with my adviser, and I want my students to do the same.”
“Although this may sound clichéd, I really think that I was born to do this. My uncle even called me ‘professor’ as a child,” he said, with a chuckle.
“I always loved teaching people and learning new things. Some of my favorite moments in class are when the students ask questions. The students challenge me, which helps me learn.”
Being challenged, either by his students or his research, is an everyday occurrence for Çatalyürek. His research focuses on multiple, overlapping elements of HPC that sit between three principal areas: parallel computing, combinatorial problems, and application.
His parallel computing research deals with optimizing software to run better in supercomputing environments. With combinatorial problems, he analyzes related components of discrete algorithms to figure out the best ways to solve problems and in what order. These two areas support Çatalyürek’s third focus, which is making his work useful to everyday people.
For example, his research is helping to predict the spread of pandemic viruses with improved results by designing and improving algorithms for software that could help to devise evacuation plans or other disease prevention activities.
“Right now, there are no good tools to show analysts how information is related, so we are looking at several different perspectives and integrating that knowledge to gain a better understanding,” Çatalyürek said. “For example, genomic sequencing is getting cheaper and cheaper to do. One day it will be affordable and advanced enough to identify specific contagions on the spot, which will help determine what areas to vaccinate, who is infected, and more.”
Another application Çatalyürek supports digitizing pathology slides. His research aims to help pathologists review potentially diseased tissue samples by building software and algorithms that create a digital version of a tissue sample. Once completed, the software will be able to scan an entire tissue sample in seconds instead of hours.
While Çatalyürek’s research is complex and touches on many different areas of supercomputing, the challenges it offers excite the veteran professor as he looks to contribute to CSE’s solutions-orientated culture.
“CSE is a highly active school with a great body of work and students, which is what attracted me here,” he said. “The culture invites collaboration, and I look forward to the challenges we will face.”
After 20 years, Çatalyürek’s life is still fueled by the discovery of new and intriguing problems. Whether those problems rest in cybersecurity or bioinformatics, Çatalyürek is driven to confront them, and he is excited to take on those problems with the newfound support of the students and professors in Georgia Tech's School of Computational Science and Engineering.