By Rob Knies
January 20, 2010 5:00 PM PT
In January 2005, as part of the inauguration of Microsoft Research India, Rick Rashid, senior vice president of Microsoft Research, outlined the value he thought the India lab could deliver. “We think research in India can result in technology solutions that can enhance the quality of everyday life for people across the world,” Rashid said at the time. A half-decade later, Microsoft Research India has taken giant strides toward validating that belief, having delivered significant contributions to its home nation and the world beyond. In advance of the Jan. 22 celebration of the lab’s fifth anniversary in Bangalore during TechVista, its annual research symposium, P. Anandan, managing director of the facility, took time to reflect on a most productive and rewarding period.
Q: What was your motivation for starting the lab in India?
Anandan: I’ve always believed that, from a research perspective and with regard to high-IQ engineering talent, India is significant. In the U.S., in all engineering and computer-science disciplines, a significant percentage of Ph.D.s and people in the leading research roles in academia and industry are of Indian origin.
It’s clear there is talent here, but often, it doesn’t get exposed and leveraged. Yet India is undergoing a great transformation, perhaps one of the greatest in its history. We are in an aspirational, optimistic, and progressive phase in our growth. Growth is founded on research and innovation. I wanted to bring Microsoft Research and its expertise and opportunities to Indian talent, and to bring Indian talent to Microsoft Research.
Q: What goals did you establish at the outset?
Anandan: We wanted to create one of the best research labs in India—and the world—in computer science. And we wanted to do basic research. There was a risk that people would see a lab in a country like India as potentially more likely to be useful for advanced development than for research. We wanted to make sure that we really became successful as a research lab, while, at the same time, leveraging opportunities for tech transfer. The goal was to make sure that ours was a first-class research lab.
Being in Microsoft Research for almost a decade helped me understand the core values we needed to emphasize and provided a model of how to achieve that. I thought, “I’ll bring that learning to this lab.”
Q: How have you fared in achieving those goals over the last five years?
Anandan: During the lab’s inauguration, Kapil Sibal, the Indian minister of Science and Technology, challenged us to be a beacon of light for research in India. I would say we are doing that. In some ways, we are a little farther along than I would have expected in the beginning. I am delighted at how many good people we have been able to attract. I feel they have measured up to world-class Microsoft Research standards. They publish in the best technical forums, and they are recognized.
I also have been impressed that many experienced researchers of Indian origin living abroad wanted to come back and help drive this. There is a lot of interest among researchers of Indian origin, both within Microsoft Research and outside, to do something like this in India. I get a lot of support from my colleagues within Microsoft Research, and it has not been difficult to hire good people.
Q: What sort of challenges have you faced in this environment, and what have you done to overcome those challenges?
Anandan: I wouldn’t say there have been any overwhelming challenge. We’ve had challenges, just as with any business getting started. One is that we don’t have enough Ph.D. production in India. We certainly don’t want to only have people who get their degrees abroad. There we have had some challenges, but we’ve managed to attract some of the best Ph.D.s coming out of India.
There is a second challenge: the perception about who we are. Inside India, people often looked at us as an R&D center, and their interaction with us would reflect that. It took a while to explain that we are not an R&D center, that we are doing research.
Abroad, there was a question about whether the India lab would be as good as the other Microsoft Research labs. I think we’ve overcome that.
Q: When you think back on five years of Microsoft Research India, what makes you smile?
Anandan: First is the quality of people we’ve been able to attract. For example, Ravi Kannan [principal researcher and head of the Algorithms and Search Group], who is at the top of his field, had been thinking about coming back to India. He had never worked outside academia, and he’s a theoretical researcher. For him to come to a place like this required a certain leap. When I first learned of his interest to come to India, we wanted to help him with whatever he wanted to do, with the idea that if he wanted to join us, we would be more than happy. Very soon, he decided to commit.
Neeraj Kayal, an outstanding young researcher, has joined our lab. He feels like he has as much freedom here as he would anywhere. That’s something that makes him very happy.
The second thing is the passion of our people to make an impact. I think about the stuff that Joseph Joy [principal software architect] has done. He really wanted to stay in the research environment but to do something that would have a real impact. We were able to create an environment where he could do that, and that defeats certain stereotypical perceptions about the barrier between research and products. I think we are able to do things that prove that point, and we are very happy about that.
We have been particularly happy about someone like Kentaro Toyama [until recently, assistant managing director of Microsoft Research India], who had no background in India. His first visit to India was when I asked him to look at whether something could be done in the space of information and communications technologies for developing countries [ICTD]. I gave him less than two weeks’ notice—I was heading here two weeks later—and he got his visa and came to India. It’s amazing to see that someone like him, from Japan and the U.S., is now seen as an important figure in India in ICTD. People have not looked at barriers of nationality and culture, and that’s very exciting.
Our lab has a very relaxed culture. One person joked that we seemed to be partying a lot! Hopefully, it’s true. I think that means that we’re all happy people. Our lab is youthful. The Indian education culture is dominated by youth at this point, and we need to reclaim that energy and deliver best practices for our young researchers to learn to use this intellectual freedom to create things. We need to maintain our culture, both socially and professionally.
When I think about our reputation with Indian academia—and I credit our External Research programs for this—a professor who spent a couple of months here said we have as good a computer-science department as anyplace. That’s not really realistic, because we don’t teach courses, but the fact that our partners would compliment us like that is a very good sign.
I have several of my old classmates from my undergraduate days who are pleasantly surprised that I could pull off something like this. When you have friends from the time you were 15, who know all your shortcomings, compliment you, it’s probably valid.
Q: What sort of high-profile projects are your researchers pursuing?
Anandan: In no particular order, I am very excited about the stuff that Sriram Rajamani [recently named the lab’s assistant managing director], B. Ashok [director of the Advanced Development and Prototyping group] and Joseph [Joy] are doing in software analytics by analyzing software much more broadly—not only the code, but all the data associated with the software-development process. There are two reasons why I’m excited about it.
One, I think it’s a game-changing thing. It is a way of helping programmers increase their productivity by bringing in not only a set of tools, but also information from past experience, in a way that will help programmers a lot more than any particular tool.
The second part is that it is an effort intended to have a direct impact on Microsoft developers. We’re working with a number of product groups, but it’s still research. It’s not like we have done something and we are packaging it to give it to the product groups. We are learning, through talking to them, what the challenges are, then researching them. That type of tight integration of impact and research makes me happy.
I’m excited about the impact and the approach we are taking in the Technology for Emerging Markets area. You cannot throw technology at these problems without thinking of the social-science aspects. People have always known that, but we’ve worked at bringing these communities together. I’m very proud about that.
Another thing I’m really excited about is our language research. It’s not an easy area in which to build significant strength in India. It’s not as if there is a unique set of problems specific to Indian languages. The problem, we learned very early on, is to get the right level of resources needed. We have a small group, but we have used it with great efficiency. The WikiBABEL project is a simple but amazing piece of work, in terms of what can be done in Indian languages, and other languages, with a small group.
I’m obviously proud of the Digital Heritage project, in which I’m involved. Not only are the concept and the technology and the research and the demos interesting, but we have been able to build a community of research in India around it. We have been able to catalyze a national research project through working with our partners.
Q: Any others?
Anandan: It’s pretty clear that the cloud has arrived and is here to stay. But it’s also clear that the story goes beyond the cloud itself. We have a project called CocoNet, which is about a layer in the cloud for content compression that delivers the operational benefits of centralization and the performance benefits of distribution.
I’m excited that Ravi is leading a movement in the theory community itself, getting algorithms research to shift its attention to contemporary problems. The community is pulling in a lot of modern ideas from heuristic approaches and trying to formalize them, and Ravi is among the leaders in that movement.
Also, our Robust Location Search project is comparable to traditional address-search engines, but it can handle misspellings and other issues. It’s getting rolled out in Bing Maps.
Q: What contributions has your External Research group made?
Anandan: One of our goals is to work with our Indian partners in academia and government to advance the state of research and research education in India, especially in computer science. We wanted to establish ourselves and gain credibility as serious and valuable partners for the science community in India.
Vidya Natampally [director of Strategy and head of External Research] and her team spent a lot of time listening to partners and understanding where the challenges are, to the point that the community feels like we are part of the overall discussion about the state of research in India. There is no “us versus them, industry versus academia” attitude toward us. They treat us like we are one of them. To the extent that Indian computer-science research continues to improve, and the number of Ph.D.s and the quality continue to increase, we will have been an integral part of that.
In the last five years, we have focused on the academic side. We feel like there’s a lot of awareness there, and the government has taken major steps to push India ahead in that area. We will be partners there. But the next place we should take our research is the industry, especially the Indian industry, to bring them to the table and come up with creative ways of building partnerships and cross-organization collaborations to enable Indian industry to invest and support more research. Ultimately, careers have to be made in the industry. I think that will be a significant focus for us in the next five years.
Q: From a personal perspective, how rewarding has it been to return to your homeland and establish a world-class research lab?
Anandan: For most of us who grew up in India in my generation, we got one of the best engineering educations you could get anywhere in the world. To find a way to do something in the country with that education … there is a very strong emotional component to that. I’m sure many of my friends and colleagues are amazed at this opportunity.
Q: What do you foresee for your lab in the next five years?
Anandan: We are at a point where we have established ourselves as having a strong group of researchers. In the next five years, I would like to see us being able to leverage it in a few areas.
Five years from now, we should look back and say that advances have occurred in the areas of software engineering, algorithms, systems, cryptography—areas where the field has evolved because of the thinking generated out of this lab. I want us to be remembered for leadership that has impacted research.
The second thing is that I would like us to have had significant collaborations with Microsoft product groups and business groups, to have done a few major things in the form of services and products. I’m not looking for them to have come out of this lab, but rather from our collaboration with the product groups.
Five years from now, externally, I hope we are regarded as a world-leading computer-science research lab. Internally, I hope our product groups say we are a most valuable partner in achieving their business and technology goals.