From 2008 until 2010, I was a graduate student at the University of Virginia. I received my MS in Computer Science, working with Kim Hazelwood. I also did some work with Wes Weimer.

With Wes, I worked on an analysis of bugs and their fixes in an effort to find defining characteristics, as well as characteristics that provide predictive power. In other words, I was attempting to discover the characteristics of a bug or its fix that make it different from other bugs, and how many characteristics there are. Does a bug that has to do with bounds checking an array have similar characteristics in the time and effort and complexity to fix as a null pointer exception? If we want to analyze a thorough sampling of bugs, what styles of bugs are important to deal with?

Kim and I are working on modifying a Pin tool that checks memory accesses. The tool checks to make sure that memory is initialized before it is accessed. However, this tool dramatically slows down the execution of the instrumented code. We are working on a potential optimization that would involve fewer checks when running through loops.