Julia Olkin is a senior research engineer for Vista Research, Inc. in Mountain View, California as of Aug. 11, 1997. Presently she is working on a project for Stanford University involving electromagnetic propagation at low grazing angles over the surface of the sea (e.g. an antenna on a ship roving the ocean). Vista Research is a small (@30 people) version of the engineering group at SRI International in Menlo Park, CA. where Julia worked for 10 years prior to joining Vista. SRI, an independent, nonprofit corporation, is one of the world’s largest research, technology development and consulting organizations, providing consulting support for government and industry.
"One of the main projects I've worked on is Active Noise and Vibration Control," Julia explains. "The idea is to use microphones or sensors to sense an unwanted noise or disturbance. These are connected to a computer which acts as the controller. It sends out anti-noise (the same disturbance but with opposite phase) through loudspeakers or actuators so that in a designated area the noise and anti-noise will sum together to zero. The controller design problem boils down to a constrained optimization problem. Using convex optimization techniques, and specially designed algorithm and code, because tens of thousands of variables are involved, we have had great success with our systems."
Work on the project is done in teams with electrical engineers, mechanical engineers, and physicists -- most have Ph.D.'s, some have masters, and a few have bachelors degrees. "Generally, the bigger projects have required everything from hardware, software, firmware, algorithms, and theory to build a successful prototype," she explains. "I was hired for my background, which is numerical analysis. When I joined up, there was a clear need to improve the algorithms designed for the controllers solving Active Noise Control Problems. It was one of those rare occasions when my thesis work was directly applicable to my job."
"Generally my role is to help people translate their physical problem into a mathematical formulation and to help solve it efficiently. Sometimes people in other fields don't realize that there might be a better way to solve a problem on the computer. In fact, they may not realize that their solutions could be incorrect! Part of my role is to educate people of the role that numerical analysis can play."
Julia has a B.A. in mathematics from Pomona College, and an M.A. and Ph.D. in mathematical sciences from Rice University. She grew up in California and originally moved back to the area for her husband's job after she completed her Ph.D in 1986. "I figured that finding a job in Silicon Valley wouldn't be too difficult," she recalls. "It was a case of networking. Prof. Gene Golub, at Stanford, has been a family friend for years. He was consulting at SRI and knew that the department needed someone with my background. He was the connection, and gave my name to SRI and vice versa. I contacted them, came in for a pre-interview, and was hired in no time."
"I was never keen on teaching, worrying about tenure, moving around, although I have always enjoyed tutoring -- the one-on-one experience. So all along I had it in mind to work in industry. During graduate school I spent summers working for Exxon Production Research and Shell Development Co. in Houston, Texas. Industry can be different from universities. In particular, the data is never 'nice.' There is always noise involved, nothing fits the mold exactly. The problems are challenging, and I enjoy the teamwork involved."
Her observations about what background is useful in industry is that "some of the hard courses turned out to be quite useful, cropping up when you'd least expect it -- such as real analysis. I thoroughly enjoyed all my numerical analysis courses and they've all come into play at one time or another. In retrospect I wish that I had taken more science courses in areas where mathematics is applied (all the areas of engineering, geophysics, physics, biology, etc)."
"It also helps to be proficient at computers, to know some of the latest languages (definitely MATLAB if you're into algorithm development), and to be comfortable using computers. A strong background in linear algebra with the basics of numerical analysis is extremely helpful."
When asked about any observations she has about being a women in industry, she says, "I've noticed a personality difference between men and women, regardless of ability. Men are much less shy about speaking up in meetings and interviews. Women are less likely to ask questions or question what has been said. So far in my career I have not noticed any problems because I'm a woman. I've been fortunate in that I work 4 days and week and still managed to get promoted and get raises."
Return to Archived Profiles and Forums
