Dr. Chandrika Kamath is a computer scientist at Lawrence Livermore National Laboratory.
Dr. Chandrika Kamath is a computer scientist at Lawrence Livermore National Laboratory (LLNL). Chandrika was born in India, and obtained her undergraduate degree in Electrical Engineering from the Indian Institute of Technology, Bombay, India. She came to the US to pursue her interests in the design and development of parallel numerical algorithms and obtained her M.S. and Ph.D from the University of Illinois at Urbana-Champaign in Computer Science. She then worked at Digital Equipment Corporation, developing mathematical libraries for high-performance systems and providing feedback to improve the hardware and compilers. A chance encounter with an engineer working on the Altavista search engine introduced her to mathematical techniques for mining text documents. On joining LLNL, she changed her field to data mining, focusing on the analysis of data sets from scientific observations, simulations, and experiments. Her current research interests include signal and image processing, machine learning, pattern recognition, and statistics, as well as the application of data mining techniques to the solution of practical problems. Chandrika holds several patents in data mining and received the R and D 100 award for the Sapphire scientific data mining software.
1) What inspired you to work in STEM?
My mother shared with me her love of mathematics and inspired me to see the beauty in numbers and patterns. I found the logic of science appealing and my father, though not in STEM, encouraged me to follow my older siblings into engineering, even though this was a rare choice for girls in India at that time. My teachers were also very supportive; I attended a girls school and we were expected to excel at whatever we did, including math and science.
2) What excites you about your work at the Energy Department?
I particularly enjoy the diverse aspects of the work I do. There is the diversity of technical topics. I work at the intersection of mathematics, computer science, and applications. So, I design algorithms for data analysis, convert them into software, and apply them to practical problems. This requires a multi-disciplinary background. It also means that I am constantly learning new things and borrowing ideas from different fields to solve a problem. I also like the diversity of problems we work on in the DOE complex. In my 17 years at LLNL, I have analyzed data from many different domains including astronomy, remote sensing, simulations of fluid mix, materials science, plasma physics, and wind energy generation. As a result, I work with people from different technical backgrounds. I find it fascinating that our different perspectives often result in innovative solutions to difficult problems.
As an example, one of the applications I am working on is additive manufacturing, where we want to build metal parts with desired properties, such as high density. However, the process is not well understood and the simulations and experiments are quite expensive. I am using data mining techniques to carefully design and analyze a small set of simulations and experiments so we can quickly determine the appropriate parameters to use in building a part.
Another project in applied mathematics involves the development of new algorithms to analyze streaming data from sensors. These data have to be analyzed as they are being acquired. For example, we want to monitor weather data near wind farms to understand if the wind energy
generation might change by a large amount in a small time. This is challenging both due to the volume of data and the speed at which they are collected. In addition, any predictions we make about the wind energy generation based on the weather data have to be accurate.
A third example is a computer science project in exascale computing. We expect that future high-performance computer systems will not be able to write out all the data from simulations. So, the current approach of first writing out the data and then analyzing them will not work. A proposed solution is to move the analysis codes into the simulation and write out only the analysis results. But, if the simulations are in support of scientific discovery, we do not know what we are looking for in the data, so we do not know which analysis algorithms would be appropriate. To address this, we are looking into intelligent ways of reducing the size of the data, without losing any of the information in them.
3) How can our country engage more women, girls, and other underrepresented groups in STEM?
This is a complex problem, but one we must address, else we risk not benefiting from the skills of a large percentage of our population. I think it would help women, girls, and under-represented groups to have mentors and role models so they do not feel isolated. As a kid, I was
inspired when I read biographies of women scientists and decided that this was something I would like to do. I find that girls often reject a STEM career based on the stereotype of a scientist or engineer. If we make them aware of the enormous variety in STEM jobs, they might
identify something that appeals to them and realize that they can use their skills and aptitude to create a rewarding career for themselves and, at the same time, make a difference.
4) Do you have tips you'd recommend for someone looking to enter your field of work?
Be curious. Don't hesitate to ask lots of questions. Be willing to learn new things. Keep an open mind - sometimes you come across an interesting project or idea when you least expect to do so. Acquire a good background in math - it isn't difficult and is used in almost all areas of STEM. Learn to write well and express your ideas clearly. Enjoy the work you do and have fun doing it.
5) When you have free time, what are your hobbies?
I like to read, cook, and garden. When time permits, I like to go down the road less traveled and visit national and state parks.