I designed (with Van Vuong) and implemented a HTML5 web application for tablets that was used by the Motorola TVEverywhere initiative. The application was designed to connect to the TVE backend and provide a listing of recorded, ongoing, live and scheduled programs which could then be played back on the tablet. (2012)

I designed this simple proof of concept web application to demonstrate technology developed in-house for converting closed caption in videos to text. The application hosted on a Nodejs server and presents the ability to outline a text box on the video. The application then sends the selection over to the media analysis backend where the text detection algorithm is executed and the text is returned and displayed on the browser. This is joint work with the media analysis team at motorola labs. (2012)

I designed and prototyped a distributed context-aware middleware system that captured and associated an individual’s predisposition to communicate with a contact (based on the history of past interactions). The system then processed these associations and enhanced the phone contacts application to display a personalized availability indicator for each contact or a set of contacts. The notion was that users could then use this indicator to initiate communication with their contacts. The concept was to distribute the application so that the mining and prediction algorithms would execute on a distributed server. (2007)

Related disclosure: System and Apparatus for Probabilistically Indicating Personalized Availability of Address Book Contacts

In 2006, I worked on a project where we designed and developed a prototype middleware framework called “Cache and Carry” that exploited predictive caching to present applications with a seemingly infinite content store, while allocating storage fairly across multiple applications. This was achieved by dynamically pre-fetching content items during periods of connectivity and garbage collecting of irrelevant (used) content based on their perceived utility. (2006-2007)

Here is a hilarious video of me trying to shoot an impromptu demo video for this project.

This project was a part of the NEST (Networked Embedded Software Technology) program at ITO, Darpa. The goal was to develop a spy location tracking system, using a sensors (Berkeley motes), iPaQs, Laptops and desktop machines. The project started out as a joint project with UCI DreamLab, but later we ended up as two separate groups programming the Location Tracking System separately.