Led the retraining of a legacy neural network for a time-series event detection task, resulting in a 20.92% improvement in validation accuracyDebugged and maintained deep learning sensor data pipelines, ensuring efficient and error-free data flow through Kafka, PSQL, and Amazon S3Engineered robust data cleaning, normalization, and augmentation scripts for sensor dataWrote scripts to visualize and evaluate neural network performance, which were used to guide technical decision-making and communicate value to clientsWorked on a team to refactor legacy code, leading to a more flexible and elegant codebase

Developed and maintained database for object detection tasks,Developed novel data augmentation techniques and model evaluation scripts,Trained, evaluated, and deployed deep neural networks for object detection and anomaly detection tasks,Researched generative deep learning techniques for project proposals,Managed digital environments in Ubuntu for various deep learning tasks