This project explores Machine Learning Operations (MLOps) as a solution to the challenges of deploying and maintaining ML models in production. It implements a flexible multi-container system that enables automated training, deployment, monitoring, and optimization of ML models. The approach is demonstrated using a distance measurement application with a rearview camera but is designed to be easily adaptable for various ML applications. The system ensures scalability, reproducibility, and maintainability, making it suitable for a wide range of ML use cases.
This project is closely related to my Bachelor's thesis, which was developed in collaboration with Bosch Engineering GmbH and Hochschule Heilbronn. The thesis explores the concept of Machine Learning Operations (MLOps) and its practical implementation. You can find the full thesis here.
