The development and productization of AI/ML solutions for smart manufacturing such as predictive maintenance, visual quality inspection, and dynamic equipment control is a challenging problem that can be solved efficiently only using a powerful infrastructure for data and model management. In this talk, we analyze several real-world case studies about implementing such infrastructures for global manufacturing companies and discuss best practices and common pitfalls. 

The attendees will learn: 

 - How to plan the development of AI/ML platforms for smart manufacturing and IoT 

 - How to design complex solutions that involves edge and cloud components 

 - How to manage sensor and image data collection in complex environments with multiple suppliers 

 - What are the limitations of the cloud-native IoT services 

We will also review several examples that illustrate how AI/ML projects can fail because of incorrect environment models, selection of inappropriate machine learning algorithms, and scalability and latency issues. We use examples mainly from the electronics and materials domains. 

This talk will be useful for process automation and AI/ML leaders who are responsible for planning and delivering smart manufacturing solutions.