Tetsuya Sakai

Why it had to be Microsoft Research Asia

Tetsuya SakaiTetsuya SakaiMost stories start at the beginning, but I’m going to break that trend by starting at the end. I am currently a Lead Researcher in Information Retrieval at Microsoft Research Asia, having arrived here in May 2009. I joined Microsoft Research Asia for many reasons, but two stand out from the crowd: the fun of being able to collaborate on a global level, not only with the wonderful researchers in Beijing, but also with those in Cambridge, Redmond and elsewhere across the world; and the overwhelming presence of Microsoft Research at research presentation venues, such as international conferences. There really is no better place to be than right here and right now.

These reasons speak to me very clearly and mirror the paths I’ve found myself traveling down during my life up to now: I’ve always wanted to be a global person. Although I’m Japanese, I actually spent five years of my childhood halfway across the world in London, learning a new language and picking up the essence of a different culture. I’ve also worked from Japan in collaboration with a U.S. venture capital company. I have always wanted to see more of the world and, being that I chose research as my career path, I’ve always wanted to work with the best. I’ve been very lucky in my life and it sometimes feels as though fortune has led me inexorably to Microsoft Research Asia.

After finishing High School, I went to Waseda University where I studied Industrial Engineering, attaining both a Bachelors and Masters in this field. Rather than go directly into a PhD course in this subject, I went straight out to work for Toshiba, where I ended up staying for 15 years. While working at Toshiba, and alongside a venture capitalist company in Boston, U.S., I helped to set up NewsWatch, which is still running successfully in Japan. In fact it was the highly enjoyable work I did on NewsWatch that helped me to attain my PhD and guided me towards my chosen research direction.

In Japan, it is possible to get a PhD while working for a company, if proof could be provided that the individual had undertaken enough high-quality research work in a certain field, and had journal papers published in that field. While setting up NewsWatch, I had managed to publish several papers and, armed with these, went back to Waseda to speak to my former supervisor, Shigeichi Hirasawa. He helped me to merge my papers to create my final thesis, and, after defending this to professors at Waseda, I was awarded a PhD in Information Retrieval and Filtering Systems.

Having just received my PhD, I was encouraged by Toshiba to go and spend time at an American university, since they attract many of the finest minds in Computer Science (with a few exceptions, of course). However, I ended up opting to become a visiting scholar at the Computer Lab at Cambridge University (one of the aforementioned exceptions) for two reasons: Having enjoyed the time I spent as a child in England, I was eager to go back and reinforce my connections with that country. Secondly, I was given the chance to work with the best people, at the time, in Information Retrieval. This included Karen Spärck Jones (who was married to Roger Needham, the former MD of Microsoft Research Cambridge) — the Yoda of IR — and Steve Robertson. These two between them invented the probabilistic model on IR, so getting the chance to work with the leading players in my chosen field was an opportunity too good to turn down. It was this experience, so soon after completing my PhD that led me to want to be a researcher.

After an excellent 18 months, I returned to my native Japan knowing what I wanted to do and after a few more years at Toshiba, I decided to act on my desires. The time spent at Cambridge had really turned me on to IR in a big way and I therefore wanted to get involved more in web search, so I left to join NewsWatch, the company I helped set up while at Toshiba.

In my second stint (of sorts) here I was both the director of Natural Language Processing (NLP) and the only researcher, which sounds a little more hectic than it was (although I was still very busy!). My research helped to set up a different kind of search engine that is still running now, called KotobaNoUchu (English translation “Galaxies of Words”). It is a unique search engine that proactively proposes new queries to the user every ten minutes for different genres, and encourages exploratory search on the aggregated search result page. Being able to undertake research in my chosen field was something that I had longed to do for a while, so the fact that some of my efforts bore fruit was amazingly rewarding.

At this time I was giving some serious thought into moving back into academia, because I wanted to spend more time in research and less in development, while at the same time doing something practical (I wasn’t asking for much, was I?!). I’d also been attending many international research conferences and it was at one of these conferences, in about 2005, that I met Chin-Yew Lin, a research manager at Microsoft Research Asia.

At conferences it’s very easy to keep bumping into the same people over and over again and I kept seeing Chin-Yew wherever I went, so naturally we got to know each other. For three years Chin-Yew badgered me to think about the possibility of joining MSRA, but I had my heart set on a return to academia and didn’t think that a business could offer what I wanted. However, it wasn’t long before I began to notice certain things about Chin-Yew and Microsoft Research Asia. For a start, it looked like he was really enjoying his work and his life (he still enjoys it, by the way), even though he wasn’t in academia either; if he was prepared to recommend his place of work to others, he must have held it in high regard. Not only that, but it was impossible to miss the fact that Microsoft Research are an ever present at these conferences; they are impossible to miss. Chin-Yew’s blatant enthusiasm for his employer and for the work he was doing struck a huge chord with me. The more he would say about MSRA, the more I began to think that it was the perfect place for me to be.

What shocks me the most now is that it took three years of persuasion before I joined Microsoft Research Asia, because looking back there could be nowhere else for me. As I’ve already mentioned, fortune has smiled on me, allowing me to consider myself a global person; Microsoft Research is the ultimate global research organization. I get to be a researcher in Information Retrieval, going to international conferences and working with the finest minds around; it’s what I’ve always wanted to do. Quite simply, it had to be Microsoft Research Asia.