Principal Software Engineer
CurrentRincon Research designs, builds, tests, and fields digital signal processing (DSP) products and services for the United States Defense and Intelligence communities.Lead architect on a modular, event driven, high data rate system where applications can be installed and be agnostic of the underlying hardware. Software lead for a project to convert an RF system to digital for a major government agency. Lead developer responsible for parallelizing and optimizing DSP algorithms.Accomplishments● Brought in to fix a project that was on its way to failure. Made major architectural changes and successfully delivered the project.● Restructured an algorithm to take advantage of CPU vectorization, achieving a 2X performance increase. This algorithm had already been heavily optimized over a 10 year period.● Implemented a phase coherent, parallel, N-stage, sub-band tuner capable of real time tuning of a 100 megasample per second input stream on a 2006-era quad core CPU.● Implemented a modular parallel correlation engine that was 300% faster than the previous implementation.● Implemented time classes used by all scientific code to support femtosecond resolution time.● Implemented a FFT class used by all scientific code to abstract away the FFT internals while simultaneously allowing for the use of high performance FFTs.● Implemented a modular TDOA calculator used by all scientific code.● Implemented a water tight binary visibility ray tracer using Optix/CUDA on dual NVIDIA GPUs.● Implemented a Lustre parallel file system to increase the I/O throughput 1000% over the previous system.● Implemented a REST API for providing multi-application access to a data store.