We would like to allow car drivers to listen to audio streams of their choice without taking too much attention off the road. To do this, we will design an application, which runs on a mobile device, that listens to an open-ended spoken request, and then plays the relevant audio stream through the cars speakers. Existing technologies such as Siri rely on natural language processing, which has limited capabilities to understand an open-ended request. In order to allow fully open-ended requests and unlimited forms of audio streams, we plan to explore a crowdsourcing approach, i.e. dispatching the request to other human beings and ask them to perform the search.

I have done a UROP with Prof. Rob Miller on Collabode, a real-time collaborative software development platform on the web. I spent my summer at Google, working on an optimization project related to embedded systems in a distributed network. I have been to international competitions at several countries in Europe and Africa.