Developed a Node.js service which when triggered, automatically seeds all client MySQL databases with the most recently released internally-built stored procedures, reducing a multi-day process to minutes

Built a listener service which recognizes events off Amazon Kinesis and forwards details to the appropriate service using Node.js

Modified a Node.js service to inject queries into pre-built MySQL stored procedures for use by other services

Wrote all unit and integration tests for developed software leveraging Mocha and Sinon.js

Developed multi-object tracking, dense-depth map image production, and feedback controller software

Performed literature reviews on state-of-the-art LiDAR-camera sensor fusion methods and object prediction methods

Configured and characterized vehicle’s LiDAR, camera, and radar sensor stack

Built a configurable recurrent neural network (RNN) optimized to execute on Groq’s machine learning chip (demoed on multivariate time-series weather data) saving days spent creating custom models for clients

Developed a model testing and benchmarking harness allowing for automatic model execution, evaluation, and dataset loading

Evaluated the performance and analyzed the structure of state-of-the-art deep learning models