Worked on LinkedIn’s social graph backend, a high-concurrency Scala application. Addressed scalability problems related to highly-connected users by changing the way connections are stored in memory and on disk.
Wrote a naive Bayes classifier for uncategorized businesses. The project made it to production, and the classifier outperformed humans from Mechanical Turk — see this article or our original paper.
We developed “azTrace,” a profiling and analysis tool which works with several languages. I wrote the memory profilers for Java and PHP, using a combination of instrumentation and reflection.
Developed web applications for local clients, which included schools and research labs.
Java, Python, Ruby, Scala, Go, C++, Bash
HTML/CSS, JavaScript, WebGL
Git, SVN, LaTeX, Berkeley DB, Postgres, MySQL, AWS
1st place for Most Useful. We developed the “Heelblazer,” a foot typing system with intelligent word prediction, using a prototype pressure-sensitive keyboard from Microsoft Research.
Competed in the Southern California regional. Placed 6th (2008), 13th (2009), 3rd (2010), and 8th (2011).
Advanced to Round 2 in 2009, 2014, 2015 and 2016.
Relevant courses: Information Security, Neural Networks, Algorithms, Data Structures, Computer Systems.