Sachin Katti, MIT PhD ‘09


Sachin Katti MIT PhD'09
Sachin Katti is currently an Assistant Professor of Electrical Engineering and Computer Science at Stanford University. He received his PhD in EECS from MIT in 2009. His research focuses on designing and building next generation high capacity wireless networks using techniques from information and coding theory. His dissertation won the 2008 ACM Doctoral Dissertation Award - Honorable Mention, the George Sprowls Award for Best Doctoral Dissertation in EECS at MIT. His work on network coding was also awarded an MIT Deshpande Center Innovation Grant, and won the 2009 William Bennett Prize for Best Paper in IEEE/ACM Transactions on Networking. He has also won the Best Demonstration Award at Mobicom 2010, Best Paper Award in ACM Homenets 2011, as well as Hoover, Packard and Terman Faculty Fellowships.

I arrived at MIT in 2003, fresh off a boat as one would say, from India. I had just finished my undergrad in EE at IIT Bombay, an extremely selective institution itself. But MIT blew me away with its intensity; the energy level of the place and the enthusiasm of the people there to engage intellectually were amazing. The first few months of fall 2003 were perhaps the most intellectually intimidating—as well as rewarding time I have ever experienced. (Imagine taking an algorithms class from an MIT Professor, Erik Demaine, who is younger than you!) The environment naturally encouraged you to explore research areas and to push yourself. It wasn’t an easygoing atmosphere; you were constantly challenged to defend your ideas. Moreover, I was part of the Systems group in CSAIL, a research group with impressive lineage dating back to the time when the original DARPA networks and TCP/IP protocols that presage the Internet were invented.  People, and especially my advisor Dina Katabi, forced together hitherto disconnected disciplines to solve the challenging problems networks face. 
The years I spent at MIT have fundamentally shaped my research taste—to ‘look for connections.’ My thesis research built an interdisciplinary approach to engineer modern wireless networks, by leveraging approaches in information theory such as network coding and applying it to design wireless mesh networks. These two areas were quite separate historically, having been shaped by an unwritten contract born decades ago.  Electrical engineers were expected to deal with messy analog signals, process them and deliver bits to the higher layers, where computer scientists could take the bits and figure out how to network them. The contract had worked quite well for a long time, and had facilitated the explosive growth of the Internet infrastructure. However, it was a poor fit for wireless networks. Shoehorning wired abstractions onto wireless networks had resulted in severe performance problems. My research questioned these long-held assumptions, and showed how an interdisciplinary approach that combines ideas from communication theory, coding theory and networking, can provide dramatic improvements in network performance.
I started down this research path itself quite serendipitously, and the story speaks to the amazing nature of research connections you can make at MIT from random conversations. My advisor (Dina Katabi) and I were attending a talk one day by Prof. Ralf Koetter who was visiting LIDS from UIUC. The talk was on network coding, a topic we had heard was gaining some traction in the information theory community, but we had no inkling it could be applied to practical network design. During the talk we realized that network coding could be applied to simplify many hard problems in designing wireless mesh networks on which Robert Morris was then actively working. From that initial connection, the research agenda grew into making network coding a fundamental building block for wireless mesh networks. 
MIT was a great place to do such research, from being able to talk with Dina Katabi, Robert Morris and Hari Balakrishnan about practical problems, to Muriel Médard on theoretical network coding ideas, there was an ideal blend of mentors that helped me take on this interdisciplinary topic. The experience also informed my decision to pursue an academic career. The freedom that academia offers to take on risky research agendas and have the time and resources to bring them to reality and make an impact is very rewarding. My advisor, Dina Katabi, and Hari Balakrishnan, were major influences and role models in making this decision, they taught by example how to balance great research, mentoring and teaching to have impact on the world in general.
After graduating from MIT, I spent a year at Berkeley with Scott Shenker and then started a faculty job at Stanford. As expected, there was a culture change, but in a good way. The closeness of the valley and the energy it brings to the Stanford campus, as well as the close working relationship you can have with cutting edge teams in industry have shaped my research. My recent research has married the fundamental approach that I learned at MIT with the opportunities at Stanford to have substantial impact on industry. My major research focus is on architecting next generation wireless networks; bringing together ideas from information theory, circuit design and software systems to engineer high capacity wireless networks. I have also recently co-founded a company to commercialize some of my recent research on radio design.
Alumni in EECS Connector 2012
Ten EECS alumni were featured in the 2012 MIT EECS Connector. Read about: Pawan Deshpande, MIT ‘06, MEng ‘07; Ted Tewksbury, MIT ‘83, SM ‘87 and PhD; Todd Coleman, MIT SM ‘02, PhD ‘05; Wendi Heinzelman, MIT SM ‘97, PhD ‘00; Sachin Katti, MIT PhD ‘09; Yehia Massoud, MIT PhD '99; Sridevi Sarma MIT PhD '06; Mark Somerville, MIT PhD ‘98; David Wentzlaff, MIT PhD ‘11; and Vanessa Wood, MIT PhD ‘11.
Read the latest issue of the MIT EECS Connector — the annual publication of the Department of Electrical Engineering and Computer Science.