Smartphones now are equipped with many sensors, like accelerometers and GPS. They are an attractive platform for tasking applications, which process sensor data continuously to determine the users context and trigger certain action based on predefined conditions (such as location-based reminders). However, developing these mobile applications today is difficult because of poor abstractions and poor programming support. My work aims to address these shortcomings through Code In The Air, a platform that simplifies the development of tasking applications for end-users. I currently focus on building a location layer for the system that is energy, computation, storage, and bandwidth efficient.

I implemented a help recommendation system at Google (2012) that assists Android users needing context-based help. I developed a high-performance benchmark suite at Intel (2011) which uses shared virtual memory on their Xeon Phi architecture. I presented my home automation and security project (2010) at an international workshop in the IEEE PerCom conference.