I am currently working as a Senior Research SDE in the Dialogue group at Microsoft Research, Montreal.
My research focuses on natural language generation for dialogue systems, evaluations of such systems and on end-to-end trainable dialogue systems.
I am also highly interested in text-based interactive image generation using Generative Adversarial Networks and multi-modal dialogue systems.
My research interests are broadly Deep Learning, Machine Learning, NLP and Computer Vision.
I received my M.Sc. in Computer Science from the University of Toronto where I worked with Prof. Ruslan Salakhutdinov and the Machine Learning Group. My thesis was titled Action Recognition and Video Description using Visual Attention. At UofT, I mainly worked on deep recurrent visual attention models for action recognition. I received my B.Tech. in Computer Science from IIT Kanpur. I did a research internship at Cornell University’s Personal Robotics Lab where I worked with Prof. Ashutosh Saxena on learning context-driven trajectory preferences for robots.
Prospective interns: If you have a strong demonstrated interest in Multi-modal Dialogue or Conditional Image Generation using GANs, send me an email on my Microsoft Email ID with the subject line starting with "[internship-application]".
Cornell University CS News, News Link, Dec 2013.
Attention, Kmart shoppers: there's a robot checker open in aisle six, Kurzweil AI, Nov 2013.
Robots trained to become less deadly, FOX News, Nov 2013.
Robots can learn to hold knives -- and not stab humans, NBC News, Nov 2013.
Cornell Researchers Help Robot Unlearn Stabby Motions With A Human Trainer, TechCrunch, Nov 2013.
IEEE Spectrum, Daily Mail (UK), CBS News, CNET, Huffington Post (UK), Gizmodo, PopSci, Slashdot Front page, ACM Technews, Discovery Channel Daily Planet, Nov 6, 2013.
The postal address given here is the mailing address for all Microsoft Research Montreal employees. Therefore, do not forget to write my name too.