Learning to Fly! I am a Software Engineer. My interests lie in application areas of Artificial Intelligence like Computer Vision, Natural Language Processing and Machine Learning. Since Dec 19th 2011, I have been working at Sportvision in Mountain View. Sportvision, the company, is responsible for adding visual 'bells and whistles' for live TV broadcasts of NFL, NHL, NBA, Nascar and more. Under the hood, its essentially a company which collects a lot of data (using sensors in a smart way) during each sporting event and then creates visual and textual displays to heighten your sports-viewing experience. I look forward to using ideas inspired by Computer Vision (and related techniques) to add robust features to their products. Prior to that, I spent a year at Honda Research Institute in Mountain View, California. At HRI, I was involved in a project automating navigation for car-like robots. The primary motivation was to use a single cheap camera and a visual marker to guide a 4-wheeled robot along a planned path. This project involved a lot of 2D image processing and projective geometry aspects to identify the marker and estimate the camera's position wrt to the marker. Software-based Controllers used the camera data for planning and sent out control commands to the actuators on the actual robot. In the winter of 2010, I graduated with a Masters degree in Computer Science from Stony Brook University, New York. Supervised by Prof. Tamara Berg, my graduate thesis involved designing Computer Vision-based algorithms for estimating aesthetics in images. This work showed one of the ways to use the vast source of information stored in high-level attributes of photographs. I spent the summer of 2010 as an intern at Bloomberg in picturesque New York City. Prior to this I spent a couple of years at Nokia's Research and Development Centre in the beautiful city of Bangalore. At Nokia, I was a part of the Web Runtimes team that developed Flash browser-plugins, in collaboration with Adobe, for Nokia's Series 60 range of smart-phones. My contribution was primarily in porting the plugin onto the Linux-based Maemo platform as an initiative to make the software cross-platform. I also blog.
News Dec 2011: Completed the advanced track of "Introduction to Artificial Intelligence" Fall 2011 course offered by Stanford's Sebastian Thrun and Peter Norvig. Dec 2011: Joined Sportvision in Mountain View as a Software Engineer. Feb 2011: My papers accepted to CVPR '11, Colorado Springs. - "High Level Describable Attributes for Predicting Aesthetics and Interestingness" - Sagnik Dhar, Vicente Ordonez, Tamara Berg - "Baby Talk: Understanding and generating image descriptions" - Girish Kulkarni, Visruth Premraj, Sagnik Dhar, Siming Li, Yejin Choi, Alexander C. Berg, Tamara Berg Dec 2010: Graduation, Stony Brook University ! |  |
