Raunk is a website that allows you to rate anything and see best of lists. The first iteration of the site was built in the Summer of 2011 with Jeremy Keeshin, David Kravitz, Daniel Smith and Zach Galant. I worked on a number of things including the list creation process, the notifications system, system administration, background processing and search, static images, and rating items.
Paperless is a code submission and review tool for the CS106 program at Stanford that Jeremy Keeshin and I built together. It supports drag and drop file submission by students. Section leaders can then review student submissions and add comments by selecting lines of the code. We even have keyboard shortcuts and markdown!
Draw on You is a Facebook app that allows you to draw on pictures of your friends in an MS paint style program and post those pictures back to Facebook. I wrote the PHP backend for fetching and posting the images.
CourseWare is a social network for course management that I helped build as a summer project with a team of 5 other students. The program was CURIS, Stanford Computer Science Undergraduate Research Internships, and we spent the summer designing and building the system. I designed the current look and feel and worked on the FAQ and notification system.
The Stanford Flipside is a one-page satirical newspaper created by Jeremy Keeshin. Jeremy and I built the puzzle system together which is in the style of sporcle.
For our CS224N (Natural Language Processing) final project my roommate and I decided to investigate satirical news and see if it could be algorithmically distinguished from real news. The problem seems pretty difficult on the surface, but, when examined, you see that real news articles tend to have coherent context while satirical articles often have totally wacky ideas thrown together. An Onion article for example mentions clowns, insemination, and Obama all in the same article. There are also the tough cases (e.g. “Adderall Receives Honorary Degree from Harvard”), but many of these are unavoidable. We built a web interface to our trained algorithm that you can try out here.