Original Link :
Click to access brochure08.pdf
I did not start out with a major in mathematics; instead, my background is in computer science. During my undergraduate years, parallel computing (using several processors simultaneously to run a large computer program faster) was an emerging and exciting new area of computer science. In graduate school, I learned that some of the most challenging problems requiring parallel computing are numerical in nature. I therefore focused my attention on applying parallel computing to numerical problems.
What did you study?
received a B.Tech. degree from the Indian Institute of Technology, New Delhi, in 1988 and a Ph.D. in 1995 from the University of Minnesota, both in Computer Science.
What do you do?
I do basic research and develop algorithms and software to solve problems in science, engineering, and optimization. Many of these problems involve simulating a physical system using a computer program, where real-world experiments might be too costly or impractical. For example, with the help of software, an automobile company can simulate a large number of crash scenarios to improve the safety of a vehicle. Real crash tests may be expensive and not accurate enough. These problems are usually so complex that they are often solved either on clusters of several computers or on supercomputers containing several processors. The algorithms and software that I develop help solve some of the underlying mathematical equations involved in these simulations efficiently on a large number of processors. The stress level and the number of hours worked fluctuate a lot; however, a typical research career affords a lot of flex time, which helps keep stress levels down even during long days.
How does Math help in the computer industry?
I believe the future of math in the computer industry lies in computer chip technology. Interestingly, just like a lot of other things, computer chips and the circuits laid out on them are also simulated to detect possible defects before a chip goes into manufacturing. In order to build faster processor chips, one needs to create larger and more complex simulations, which in turn, requires faster processors.