My main focus here has been working with multiple teams to improve Semgrep's program analysis in various ways, including adding support for more languages, reducing false positives in our scan results, and making our analysis run faster. The people here are amazing, and this has been one of my favorite industry experiences throughout my career.

I was on the team that owned pprof, the profiling tool. Our team pretty much served continuous profiling to most of Google's internal services. The work there included adding new functionality to the internal profiling tool, improving the deployment and release processes of our internal services, leading internal workshops on our profiling tools, and maintaining the Google Cloud profiling agent open source libraries.

Professor David Padua was my advisor and we worked on several experiments, including formal methods to help compile math down to C code, speeding up map reduce programs, writing a compiler for a new ML accelerator, and finally my thesis which is automatically generating performance tests for compilers.

Helped write the compiler for one of Huawei's prototype AI accelerator chips.