ML performance engineering

Pre-silicon development for next generation systems

Optimize Machine learning SDK subsystems, Design and implement various submodules. Using CUDA framework for GPU acceleration of SDK components when possible. Regularly conduct performance analysis to identify and remove performance bottlenecks. Improved object detection visualization performance 2X. Improved object Tracker processing speed by 1.5X. Design and source key hardware infrastructure.

Assisting in the development of computer vision machine learning SDK developed in C++. Standardized performance tests by creating company wide xml format for ground truth data. Implemented encryption routines for key configuration components. Current responsibilities include Maintaining and operating automated test server, setting up remote benchmarking tool for clients, and integrating various SDK components.