A machine learning web application built on Node.js using the Clarifai Deep Learning API. The application scrapes data off of the top hits on Google Image searches to determine what Pokemon the user resembles the most.
Best Use of Clarifai API Winner at HackMIT, 9/20/2015

A app for crowdsourcing after-school activities in underprivileged schools.
Views include a student view that includes upvoting, suggesting, and viewing event content, and a teacher view that includes sponsoring, scheduling, and adding content from an event. Built as part of the Code For A Cause initiative with Yang Dai.

A Tinder-style shopping app for Android. Users connect through Facebook and like/dislike randomly generated Amazon products and then get matched with their Facebook friends based on liking many similar items (still in progress).

A UI design for an app built on top of Uber's existing framework that uses a heatmap of varying intensities to display regions with fewer or greater numbers of potential riders based on historical data. Built in Spring 2015 as part of the UI/UX course at MIT.
The platform is built for web browsers, but the interface itself is meant to mirror that of a mobile platform. The data itself is also hardcoded (i.e. there is no backend), because the goal of this project was to focus on creating a unique user interface and user experience.

A platform for eye-based collaborative problem-solving using eye tracking data, built for the Tobii eyeX Tracker using Node.js and Python scripts. Built as part of a semester-long undergraduate research project (UROP) with Scott Greenwald for the Fluid Interfaces group at the MIT Media Lab.
Demos for this include lytro image focus based and mouse movement control based on eye gaze position.

A debate platform using a timed chat queue to moderate discussion. Built on Meteor for MIT's 6.470 Web Programming Competition 2015 with Jennifer Zhang and Weilian Chu.