•Designed and deployed data pipelines and machine learning algorithms (CARTs, XGBoost, Logistic Regression, Neural Networks) to solve problems, including in manufacturing and driver assistance, for many of Bosch’s business units .
•Built an image classification framework using TensorFlow 2.0 inside Docker, allowing for modular inclusion of pretrained CNN architectures for transfer learning.
•Implemented a bayesian updating system to integrate machine learning model outputs over time to improve prediction as measured by a custom loss function.
•Designed and coded a distributed hyperparameter optimization framework synchronized via a SQLite database.
