Professor to CEO and back again: Q & A with Dr. Vikram Jandhyala
By Eran Moore Rea
Vikram Jandhyala knows success in two worlds: business and academia.
In 2006, he took a leave of absence from his professorship at UW to lead his start-up company, Nimbic. He returned to UW in 2009 and he has chaired the UW Electrical Engineering department since 2011. He now directs the UW Department of Electrical Engineering’s Applied Computational Engineering (ACE) Lab. He also co-directs the Northwest Institute for Advanced Computing—a joint effort from the UW and the Department of Energy’s Pacific Northwest National Lab (PNNL).
Jandhyala’s company, Nimbic, provides electromagnetic simulation and design* to leading companies: Texas Instruments, Toshiba, and Qualcomm, to name a few. Learn more about Nimbic at http://www.nimbic.com/.
Dr. Jandhyala, so great to speak with you. How did you get your start in academia?
I lived in Delhi for 21 years of my life. I went to the Indian Institute of Technology, the top engineering school in India. It’s on the MIT model; the teaching model there is very similar, and it’s a very competitive environment. I was always into math and physics, coding and computation; I was just not that interested in lab work. I wanted to do something with computation, something that combined applied math and physics.
And you always wanted to be a professor? Did you want to go into business as well?
I did always know I wanted to go into academia. Both my parents were physics professors. They work on really deep stuff—biophysics, particle physics, complex systems—that has very deep theory and eventually leads to applications, but isn’t going to be ready for commercialization right away.
So I really had no background with commercial applications going into my career. But I found a lot of value, a lot of important problems waiting to be solved in industry.
And what did you think of commercialization at that time?
Well, none of us were thinking, “Oh, let’s create a start-up.” It wasn’t that kind of culture at the time.
But, almost all of our research—especially in the lab I was in—had real, commercial applications. We helped electronics designers in industry and defense improve their designs for aircrafts and satellites.
So even if we weren’t thinking of starting businesses ourselves, we were definitely working on those sorts of problems, real-world problems.
You moved from India to Illinois for graduate school. What was that like?
Yes, it was a big culture shock. In 1993, I went from 120 degrees Fahrenheit to minus 60 Fahrenheit (with wind chill, but still, cold)!
I didn’t know anyone in Illinois, but when I got there I found a wonderfully sharp advisor. I was actually his first PhD student. Dr. Eric Michielssen, he’s now a very successful professor at Michigan.
And after grad school in Illinois, you went into industry.
Yes, at Ansoft Corporation in Pittsburgh. I was a software developer there.
The job had an academic feel, though—it was almost like being a super-postdoc. A professor from Carnegie Mellon, Dr. Zoltan Cendes, started the company.
When I first got there, I almost never talked to anyone except other developers and maybe the people testing the code. But over time, as our product started becoming more popular, we started expanding—talking to everyone.
To me, that was very interesting. I began to see the impact of what I was doing.
And from Ansoft you came to the UW.
It’s all because of Leung Tsang. He was a professor in UW EE department at that time (he later became chair of the department). He knew me from Ansoft, and he told me “We’ve got a faculty position opening up in electromagnetics. I don’t know when we’ll have another one.”
So I applied, and I was lucky enough to come to UW.
How did you discover the algorithms that would eventually become Nimbic?
I was very lucky to have really spectacular students; my first few grad students were very, very brilliant. I got some funding early on from the National Science Foundation Faculty Early Career Development (NSF CAREER) grant, which is a big deal for junior faculty. That led to a joint grant with Defense Advanced Research Project Agency (DARPA).
At some stage, DARPA asked us if we had a company. They wanted to continue our funding in a Small Business Innovation Research (SBIR) grant. We, of course, didn’t have a company at that time, but it got us thinking.
Dr. Luciana Simoncini from the Washington Research Foundation (WRF) asked my department for interesting projects and the chair at the time, David Allsot, mentioned my name. Luciana suggested I work with the C4C.
And that’s the stage when we started to think about protecting intellectual property (IP), writing patents, and copyrighting our code. Fred Holt from the C4C introduced us to several Venture Capitalists (VCs), and two of them—the Madrona Venture Group and the WRF—decided to fund us.
Madrona brought in a business person, Bala Vishwanath, and we put our team together. I got two of my best students back from industry to help me start this thing, along with two other very good students who were finishing up their graduate work. We had a team of six people at the beginning.
How did you grow your company?
We started working very early with Texas Instruments (TI) and Toshiba. Right off the bat—they were not paying us anything, but they were interested in the tech and they said, “Okay, we’ll work with you to build out the tech and see where this goes.” We launched in 2006, calling ourselves Physware.
And then the cloud idea became very interesting. The cloud, as you know, takes all the software you used to download or get off a CD and makes it remotely accessible, from any device. Now, any company you think of, whether it’s Netflix or Salesforce or Google—it’s all running on the cloud. But that wasn’t totally the case in 2009, when we launched into the cloud. Even now, most companies in the electronic design industry aren’t running their enterprise software on the cloud.
I liked the name Physware, because I was very proud of the fact that this was physics software. But it’s really kind of a geeky, nerdy name. We changed our name to Nimbic when I came back to UW in 2009. The cloud is something cool and new, so it’s Nimbic for nimbus clouds (nimb-) and integrated circuits (-ic). Nimb + ic. Nimbic.
You began your leave in 2006.
Yes. I first took the Chief Technology Officer (CTO) role, but by late 2007 I realized that this was a pretty technical sale. As the CTO, I led the overall technical development of our software, but I didn’t help sell it. It became clear that the best way to sell such a technical thing was to put me, the founder, in charge of the whole company.
So I took over the CEO role. Bala, the business person, became the president. I got to talk with customers and work on the technology.
I came back to UW in 2009; I was always going to come back. At some point you get big enough to attract a big-name CEO. Bala brought Dr. Raul Camposano to our company. Raul took over when I left; he was the former CTO of Synopsys the best-known company in our field.
You said it was a very technical sale?
Yes, and no. It’s a balance.
You can’t get caught in the academic model. You can’t sell your technology by trying to prove exactly how smart you are. I had to realize that we couldn’t sell the brilliance of our algorithms. We could only sell what those algorithms can do.
But, you also can’t dumb down your science. If you simplify your science too much, customers might think you’re selling a very simple solution. They might think, “Okay, this is not a big deal.”
In fact, this was a quote from an actual customer: “Well, if you’re able to solve the problem so fast, that means it’s not a difficult problem in the first place. So why should I pay so much for this software?”
They didn’t realize, we’d taken a really complex problem and created a new way of solving it—that’s why it’s so fast.
So what did you do?
We moved on to other customers. That was something we learned early on: to choose our customers carefully.
What surprised you about being a CEO?
I learned how much marketing matters, using the right words. You have to create the right idea or feeling in people’s minds or you’re not going anywhere. That was something I just could not believe at first.
I thought like a scientist, that there’s only one truth and only one way to explain things, and everyone should get it right away. But that’s not true.
You might have the exact same technology as someone else, but if they say it better, you’re going to lose. That’s important.
So is that advertising, or talking to customers?
Both really. Communication with customers is key. As I said, sometimes a business needs to go find other customers if the communication doesn’t work out.
But I also found that sometimes customers showed me different aspects of our product—stuff I hadn’t really thought about.
For example, we had one customer who was really happy with everything . . . except that one outline of a box on one of our 3D images was bright white, not bright yellow. This guy wasn’t concerned about the speed, or the accuracy, or what it was doing. He just wanted it to be a different color.
So we changed the color. But I found out later that there was a legitimate reason he wanted it changed. This guy did a lot of design, looked at hundreds of 3D objects every day, and there’s a certain way these outlines were usually colored. So I saw it was actually a non-trivial thing from his point of view.
And there we were, the developers, so proud of our deep algorithm and our computation that we couldn’t see that.
Did your focus change after you came back to UW?
Yes, I changed my research completely when I came back. It’s almost like doing a second PhD; I took a summer off to read about new areas, and my students are all in new areas.
Now, my lab is taking the physics of electromagnetics and applying it to the kind of graphs used in social networks and cyber security. We’re trying to understand patterns of influence and information using wave propagation. We’re looking to model how people make decisions in small and large groups
I try to think about research with three goals in mind:
- Is it important enough to make impact?
- Is it in my skill set? Or, can I learn the skills fast enough given my background?
- Is it interesting enough to keep me and my students motivated and working hard?
I’m increasingly looking at all of the research my lab does through a start-up lens.
Finally, what advice do you give other researchers thinking about starting their own companies?
I would advise people to do it early in their careers when they don’t have other responsibilities. When you’re ending grad school, that’s the time to do it.
It’s not something you can do on the side, basically. It will take up all your energy.
Some academics don’t think like this. They just want their place on the board and they don’t want to lead their companies. And that’s where I think differently; if you want to do this at least once as a faculty member, you need to go the whole way: be on the founding team, CTO or CEO.
And then once you’re learned, you can advise your students how to do it. I mean, honestly, if you don’t do it yourself, you’re really missing an experience.
*Electromagnetic simulation takes a product’s parameters—a certain kind of wire, a certain design for a cell phone—and computes the real-world output in energy, radiation, speed, etc. So, if a company wants a smartphone that can run a camera, a phone, Wi-fi, games, email, internet, and text messages all at the same time without burning down the battery in ten minutes, that company needs to run many electromagnetic simulations.
Electromagnetic design takes ideal real-world outputs—how long a smartphone’s battery should last, for example—and computes the ideal design a company should use to obtain those outputs.