• Trained customers to use DeepSig software to train a neural network to detect and identify signals of interest, eventually developing model used in customer product

• Added automatic recording feature, previously only accessible via CLI, to GUI by working with with Sr. developer to refactor handling of recording parameters so they could be altered during runtime, creating getter and setter API calls for those parameters, and adding frontend features to access API calls

• Ported sections of existing backend C++ code from CPU to GPU using CUDA libraries to parallelize code, improving speed of software on devices with multiple GPUs

• Improved default model accuracy by adding “cut out” data augmentation to training pipeline of neural network, by setting random continuous groups of data samples to 0

• Reduced technical debt by identifying underlying issues, implementing solutions, and creating unit tests

• Performed thorough code reviews by checking functionality, readability and style on merge requests

• Completed QA tests each release cycle on Nvidia Jetson Xavier developer kits