Using Containers to deploy AI Models across our microscopy platform

Deploying AI models that solve computer vision tasks across platforms in an efficient and reproducible way is essential when your goal is to create systems for quantitative microscopy.

In addition the integration of the latest AI technology in bigger software platforms has to be done in a modular way that allows updating and improving the tools even for local users.

Faced with all those challenges we at ZEISS decided to try out a new way to deploy our AI models by “packaging” them together with all required code dependencies into containers, which can be deployed across all required platforms at scale while ensuring reproducible results.