I have a love and expertise for optimization, and a particular interest in CUDA. I started working in CUDA a few years ago, mostly offloading bottlenecks in matlab code written by grad students into CUDA for better performance (my record is 2.5 hours run time down to 12ms, for 750,000x speedup). I also like to improve algorithms for more accurate results, such as extending the NVidia HSOpticalFlow sample code to use color images instead of grayscale.

I started programming for real in C, with an ambitious first project. I built a TCP/IP high precision data acquisition board for strain gauge transducers based on a PIC32 micro-controller and an AD7192 sigma-delta ADC chip. I built it independently (soldering and all) as a hobby project at minimal cost. It works beautifully and is more accurate than anything i could find online, including the Agilent system the company i worked for at the time was using. I built some world class multi-component strain gauge transducers while I worked for that company and i needed an ADC to match the precision. I left the company to follow my wife for medical school shortly after I finished the prototype, and I never tried to commercialize it.

I went to school at the University of Minnesota for aerospace engineering. I worked on some really awesome projects there including leading the solar vehicle project and the formula SAE heavy lift competition (mostly first place wins, a couple second place). At the time I was primarily an aerodynamics and structures engineer. I've always liked programming, first TI-Basic in grade school, then python in highschool, then matlab in college. It wasn't until I started programming high performance applications in CUDA that I realized this is what I wanted to do with my life.