My master’s project is Sparse Subspace Clustering with Missing Entries, particularly on developing efficient and effective algorithms to be able to cluster high-dimensional data with missing entries. This project has involved developing new algorithms and theory on Sparse Subspace Clustering.
An easier way of explaining this is as follows: a service such as Netflix has plenty of users and movies, but obviously not every user has seen every movie. However, given that some movies and users are similar in certain qualities (ie comedy movies, or a taste in Cristopher Nolan films for users), one may infer that one could come up with recommendations. Hence, simulatenously group similar movies and users is exactly an application of this research.
Starting Fall 2016, I began as a piano minor at the Peabody Conservatory at the Johns Hopkins University. I learned and performed:
Video recording of one of my performances during Spring 2017:
Many years before, I produced a video recording of Beethoven’s Moonlight Sonata. Now I think it feels a bit rushed and amateurish, but still believe it’s enjoyable.
As part of my Biomedical Engineering Design Team, I cofounded flapp, a mobile application for 3D local soft-tissue flap Design for facial reconstructive surgery. Flapp established collaborations with the Johns Hopkins School of Medicine and Center for Bioengineering Innovation and Design. Further, we won several honors and grants: Finalist for JHU Business Plan Competition, inHealth Shark Tank, and the Dr. Andrew Malinow Bridge Funding Award.
What if the most creative minds channeled their focus into solving the most impactful problems of today? Imagine if we could apply the ingenuity that powers the most profound technology into the most fundamental of all human concerns: health. MedHacks is the start.
I founded and led the creation of the first MedHacks, and currently advise as executive chairman. MedHacks 1.0 attracted over 400 participants from 25 universities and 3 countries to hack health Oct 2-4, 2015. Due to our success, I was invited to speak at Hackcon IV and gave talk on digital health and medtech innovation. Now in its third year of existence, MedHacks has established relationships with numerous incubators, tech companies, academics centers, and VC firms. We’re always looking for great organizations to partner with - if you’d like to get in touch shoot me an email at [email protected] :).
Winning the Comcast EveryBlock API prize, StrollSafe is an app all about contextual maps. Strollsafe releases location-based crime reports for people out walking. Using OpenDataPhilly and EveryBlock info, the app overlays crime data onto a standard map.
Worked with a classmate in Computer Vision course to colorize Sergei Mikhailovich Prokudin-Gorskii’s black and white images of the Russian empire. Gorskii took 3 shots of each scene - one with a Blue filter in front of the camera, one with a Green filter, and one with a Red filter. Our solution used SIFT descriptors to define homographies between the 3 sub-images and merge them together in mosaic-like fashion.
Further information (code, images, and report) available on Github.
Placing as the top JHU team and 3rd overall at Hophacks Spring 2015, Dash is an incentive based fitness app that allows users to competitively race their friends. Users can schedule their races within the Dash interface, which allows them to set an endpoint and the amount each competitor must wager to enter the race. It’s simple: first person to reach the end marker wins the prize. We wanted to revolutionize the way people think about fitness, and Dash provides motivation and childlike wonder to running again. Along with being a primary developer on the project, I designed the entire UI/UX. Check out our submission on DevPost and our presentation on YouTube.