Enhance and improve GPU Image Signal Processing pipeline in C++ / CUDA / Multithreading

▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬

• Achieved higher performance on Image Signal Processing pipeline using Nvidia's GTX980/1070 in C++ and GPU programming

• Developed DLL to visualize and capture image from MIPI and to communicate withCMOS sensor through I2C under .NET 4.5.2 in C++ and Python

• Developed ISP algorithm of Auto Exposure, Contrast Enhancement, Sharpening, WhiteBalancing, Demosaic, Noise Reduction, Tone mapping, and FFT in GPU programming

• Designed the GUI to analyze, validate, and adjust image quality under Visual Studio 2015

Developed, enhanced, and benchmarked the Datapath program in C# / CUDA / multithreading / network programing under Reflective E-Beam Lithography Group

▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬

• Achieved high throughput by developing and enhancing the renderer program in CUDA and C# on GeForce8800, GTX465, GTX580, Tesla2075, and GTX780 Ti

• Satisfied customer's needs by collaborating firmware engineer to build user-friendly GUI of chip verification

• Improved performance by using ANTS profiler, visual profiler, and CUDA Nsight

• Enhanced and maintained the encoder of rendered image in C# and CUDA program.

• Maintained the driver of FPGA encoder card in C# and Jungo WinDriver