The optimization of complex systems is crucial for lower costs and greater efficiency in the supply chain. Traditional advanced process controls are often expensive and time-consuming. Machine learning can significantly reduce this effort: A model as a digital twin simulates the system and optimizes its settings. After the proof of concept, the focus is on fully automating and scaling the infrastructure and data pipeline.
The presentation will show the project's cloud migration and explain the key challenges and solutions for setting up a scalable MLOps process.