• Implemented Resnet in Pytorch, performed hyperparameter tuning, and preprocessed datasets, for a 175% accuracy increase in deep learned inertial odometry• Expanded dataset by augmenting and preprocessing 8 public datasets, leading to an additional 56% improvement in accuracy (total of 330% accuracy improvement)• Implemented real time inference code for embedded system with 17hz inference frequency• Visualized Resnet convolution state to interpret the model’s learning

• Key team member of emerging start-up developing martial arts competition management platform• Positioned to deploy our product at the annual Collegiate Wushu and SIMAC tournaments• Developing front-end utilizing React and Typescript• Worked closely with a compact team of UX designers and back-end developers

Streamlined printer information collection from 50+ APIs for HP developers to reduce triage time• Designed and Engineered a full-stack web application using React and Django to analyze printer test information• Engaged with international co-workers within a Scrum framework