I'm a PhD student in MIT
CSAIL, working with Professor Hari
Balakrishnan and Professor Sam
Madden.
I graduated from IIT Madras in July 2005 with a B.Tech in Computer Science, and received my Masters from MIT in September 2007. My research interests are in sensor networks, applied machine learning and data management.
I am currently working on the exciting problem of crowdsourcing traffic. The problem is to estimate and predict future road traffic conditions from location data gathered from people's mobile phones in real-time as they drive around a city. Specifically, I am looking at three research questions: how to obtain accurate travel time estimates from noisy WiFi and cellular location data, which is much cheaper energy-wise than GPS but much less accurate; how to use machine learning to predict future traffic conditions by looking at historical data, and to learn spatio-temporal correlations (for example, correlations between traffic conditions on adjacent roads at a major intersection), and how to optimize energy usage on the fly and save power by switching to cheaper sensors than GPS (like WiFi and cellular localization) whenever appropriate. This research is part of Cartel, a telematics system at MIT which collects location and sensor data from a deployment of taxicabs in the Boston area.
Previously, I have worked on FunctionDB, a relational database system that allows users to query mathematical functions as first-class citizens. This is a useful abstraction for many applications, such as sensor networks, which need to deal with noisy, error-prone and potentially incomplete data which is a continuous function of time or space. This is in contrast to traditional relational data, which typically consists of discrete rows or tuples. This is a logical extension of the MauveDB project.
I have also worked on Wavescope, a platform for high data rate sensor computing and streaming applications and helped build XStream, a high data-rate signal processing system which is part of Wavescope.
I have also worked on systems research projects during my summer internships at Yahoo! Research (Summer 2008) and Google (Summer 2007). More details forthcoming.
I am an avid follower, rasika and beginning student of Carnatic Music, the classical music of South India, an amazingly rich and diverse art form. Here's a web page devoted to the late Shri Ariyakudi Ramanuja Iyengar, one of my all-time favourite musicians in this style.
I am also interested in speaking Samskritam (Sanskrit), one of the oldest and richest languages of India and the world. We have a student group, MIT Samskritam dedicated to promoting and speaking this wonderful language, and we meet weekly on Wednesdays for spoken Sanskrit classes. I currently help teach the beginner class for this group.
I've started dabbling with photography a little; check out my picasa photostream here. Pics taken using a Nikon D40.