|
|
 |
» |
|
| |
By Jamie Beckett, October 2007 |
|
Prith Banerjee is not a man who likes to waste time. At 47, the new director of HP Labs has already founded two companies, published hundreds of papers, received several patents and has been honored by three of the IT industry's most prestigious technical societies.
So it wasn't surprising when, just days after he joined HP in August, he plunged into an ambitious project involving 120 people worldwide to develop a new strategic plan for the lab. That plan is to begin rolling out in November.
"I want HP Labs to do more research that impacts people and society, and I want to see this transferred to HP businesses," Banerjee says.
Banerjee, who plays the sitar for relaxation, has not had much time for his instrument lately. In the interview below, the former dean of the College of Engineering at the University of Illinois at Chicago talks about what lured him away from academia to tackle the challenges faced by HP Labs and why he believes open innovation is essential for the future.
He also shares some personal reflections about his proudest accomplishments, his key influences and his first invention.
| |
You've been at HP Labs about three months. What are your early impressions? |
|
|
|
I knew it was a world-class organization. I've been familiar with the lab for some 20 years, and already knew several researchers. After I joined, I was even more positively impressed. The quality of researchers in HP Labs is very, very high. The people are really willing. I think they are looking for some leadership to help them reach the next level.
I've been meeting with researchers in their cubicles, just getting to know them one-on-one and learning more about their work. There's tremendous excitement among the researchers and I'm happy to be part of it.
| |
What big challenges and opportunities do you see ahead? |
|
|
HP Labs, as an organization, has not been as successful with technology transfer as it should be within the context of a corporate research lab. I want the lab to do more research that impacts people and society, and I want to see this transferred to HP businesses.
One challenge I'm facing is finding the right balance between short-term and long-term research. That doesn't mean I want to swing the pendulum way towards short term, because then you run the risk of not being able to create really new business opportunities in the future.
Fortunately, we have a true breadth of experience in researchers. As a company, HP is in a wide range of products and services – you name an area – and there is an expert in HP Labs who knows a lot about it. So the opportunity is that talent exists, and the challenge is how do you get all the right talent to work together on larger collaborative projects, so that you can do high-impact research.
| |
You've talked a lot about open innovation. What do you mean by that? |
|
|
Open innovation recognizes that there are people outside your own organization who are as smart as you, or perhaps even smarter. In the past, research labs focused on closed innovation – they did the work on their own and then tried to funnel it to the businesses inside their own companies. The idea behind open innovation is that by working with others – PhD students at universities, entrepreneurs, startups, etc. – you can tap into their ideas.
But with open innovation comes open output as well. I'd like to bring some startup DNA into HP Labs, so if there isn't a place in an HP business for an innovation, researchers can start a company or HP can license the technology.
It's better to do that than keep technologies hidden inside HP Labs. As long as the research done within HP Labs ultimately benefits society, you have done something, right? Society benefits and HP as a company would get some licensing revenue or some equity share in a company.
| |
Then how would you measure your return on investment? |
|
|
I'd start with making sure that when a research project begins there's a reason why – not just a scientific reason, but a market reason, why this work should proceed.
People will need to explain in their research proposal what the possible impact could be – some rough estimate of the size of the market, how much they think HP will have to invest to bring that research to market, and what the predicted return on investment might be.
So one way we'll decide whether to fund projects is to determine if they can have a big impact. I want HP Labs to work on big-time, risky projects. Not all of these will succeed, but if there are a few home runs, that would create enormous value for HP.
| |
What persuaded you to leave academia and join HP Labs? |
|
|
I was aware of the challenges facing corporate research labs in general – the struggle balancing long-term and short-term research, the difficulties with technology transfer. And I always wondered if I was in a position to influence the direction of a research organization, the challenge would be fantastic.
It's really the challenge that attracted me. I am here to impact the way HP Labs operates, and as you can see, I literally hit the ground running. I need not have started strategic plan in the tenth day of my stay here. But I got attracted because of the challenge of making an impact in a truly world-class research organization.
| |
Over the course of your career, are there any specific projects or accomplishments that make you particularly proud? |
|
|
I am most proud of the research on MATCH (MATlab Compiler for Heterogeneous adaptive computing systems) at Northwestern University. It is, in a way, a classic story. It had all the right things. It had journal papers, conference papers and patents – and it was really advancing the state of the art.
It was one of those large, collaborative projects funded by DARPA at Northwestern. As part of that, I graduated about six PhD students from that project itself. I've got at least 20 publications from that one project and I received two patents for the work. The technology was eventually licensed to AccelChip, a startup I founded. AccelChip created a product, and then was acquired by Xilinx in 2006. Now some 20,000 customers of Xilinx use the AccelChip tool.
The key technology was translating MATlab programs so they could be compiled into hardware. People said it couldn't be done, so it was a very interesting problem impacting a large market.
| |
What was the toughest part of founding a startup? |
|
|
Staying focused and staying motivated. There are some real ups and downs in running a startup. There are days when the pressure is incredibly high. Maybe you're trying to raise the next round of funding, and people around you worry that if you can't, they'll be out of a job. Periodically the morale of the team goes down. It's difficult to keep up that level of intensity for a long time.
It is difficult on your family, too. The founder sets the pace for the company. Unless you are willing to put in those long hours, the people following you will not put in long hours. There were times when my wife and family wanted to go to a social gathering and instead I went to the office on a Sunday. I could have easily telecommuted, but I made it a point to show up because if I showed up, everybody else would show up.
I remember one day at AccelChip when the air conditioner broke down. It was the middle of summer, very hot, and the air conditioner was broken. So I told my people, "Hang in there. I'll be right back." I went to Target, bought four or five fans and lugged them all the stairs myself. It was just a little thing, but it's important to show you care for your people.
I think having a sense of community in a company is important so people feel it is more than just a place to work. That can help you attract and retain employees long-term.
| |
Can you talk a bit more about yourself? What first sparked your interest in science and technology? |
|
|
I have always been interested in electronics, from a very young kid. Now, in India, we didn't have too much, but whatever we did have -- radios, televisions, transistors, etc. -- I used to take them apart and play with them. So very early on, I knew I would be a sort of electronics kind of engineer, that it just attracted me.
I was also lucky enough to have a sort of mentor – my older brother – who was also interested in these things. As it turns out, he and I have had very similar careers. He also has a PhD in electrical engineering. He is now a Professor of Electrical Engineering at the University of Texas at Austin, where he is also director of the university's microelectronics center.
Coincidentally, we not only have parallel careers – we also have very similar academic records. We both went to the Indian Institute of Technology at Kharagpur. We both did our PhDs at the University of Illinois at Urbana-Champaign, and we both ended up in academia.
| |
Do you remember inventing anything as a boy? |
|
|
In high school, I created an automatic flow meter to measure the rate at which liquid flows through a pipe. I invented a gadget that would measure that through applications of electromagnetic and hydraulic principles.
I used to read popular science magazines and articles, and so I wanted to attack a real problem we had at home. I forget the details, but we had some issue about water flowing and clogging in our pipes in our home, and we wanted to measure it. My device worked so I submitted it as a science project and it won.
| |
I understand that you enjoy classical Indian music and that you even play the sitar. |
|
|
Yes, I've been playing the sitar for 25 years. That's my relaxation. I also listen to a lot of classical Indian music, attend concerts, all that.
My favorite musician is Nikhil Banerjee (no relation to Prith, Banerjee was one of India's most prominent sitar players of the 20th century). Ravi Shankar popularized that instrument, but technically Banerjee is superior.
I got started playing the sitar because of my father. He used to play, and we had an old sitar at home. We all used to listen to classical music, and one day I saw that old sitar and thought, "I want to learn to play that." My father was my first teacher.
| |
Many people like you who have a background in mathematics or engineering are also highly musical. Why do you think that is? |
|
|
Music, the highest quality music, is about rhythms and ratios and so on, and people who are good at math can pick up those things much more intuitively.
When you're playing Beethoven or Mozart or other classical western music, you just try to play the music following the script you've been given. But Indian music, by design, allows you more freedom. There is a structure, but you can take some exits or do something else and then come back to it. So essentially, you are composing music on the fly.
I think a mathematical background allows you to understand the ‘structure’ of that music – in Indian music it's called the raga. You have to stay within the structure, but then you can play variations on that.
|
|
|
Printable version
