Adding the appropriate tags to content (aka classification) can be a really tricky task, especially if you have more than one content manager. Our team at JSI observed this while working on the Infant and Young Child Feeding (IYCF) Image Bank, a website that serves as a repository of open-source illustrations designed to be used for international health and nutrition programs. In this session, we will show how we built an integrated machine learning model (based on Tensorflow.js) to analyze and categorize newly uploaded images according to our own previous categorizations of those images. This reduces the time it takes our content managers to add content to the site and enhances accuracy/consistency of image classification. The session will frame the issue and then do a deep dive into the code required for our solution.