Managing and optimizing yield through data analytics is an undeniable requirement in the digital transformation of the semiconductor industry. Traditionally, the tools used for analysis are locally deployed. Transferring data to the cloud and doing and analysis in cloud allows us to take advantage of cloud services at scale and keep compute cost of ownership at a minimum.
This presentation outlines a practical approach to deploying commercial applications that use AI/ML in cloud. We will talk about the benefits of cloud adoption as well how the various cloud deployment factors can be approached and addressed in context of different customer cloud adoption strategies and what to look for in a vendor solution. Drawing on best practices and our experience and purposeful product stack design we will present a few different ways of deployment and effective resource usage while also addressing the different factors that need to be considered when deploying commercial application to the cloud.