Currently working on this fun project that lets a user draw a number on an html drawing pad (using LiterallyCanvas). The drawing will be saved as an image, base64 encoded and sent to an aws lambda endpoint that will decode the image and classify the number, returning a response to the front end. I've taken a few machine learning courses but never implemented a model that is used in any real setting. I decided to work on this to see what it is like to have a model out in production (even if for useless tasks).