United States
Implemented Loop Invariant Code Motion for Virgil Compiler Implementing Unboxing and packing techniques for Virgil
* Spearheaded the SLAM (Simultaneous Localization and Mapping) algorithm pipeline and codebase for an FSAE autonomous race car. * Modularized functionality using C++ namespaces, header files, classes, and structs for organization, ease of debugging, readability, and future plug-and-play use of SLAM models. * Multithreaded SLAM pipeline and utilized vectorization techniques, achieving a tenfold reduction in computation time for performing sufficient updates to map racetrack. * Maximized productivity by advising, guiding, and debugging with subteam members through weekly meetings and work sessions.
Wrote scripts to process Python benchmark scripts for different Natural Language Processing models. Benchmarks were passed to different language compilers, and the newly compiled benchmarks were tested on the models.