I only started learning/implementing ml very recently. Gave up halfway through my previous attempt when i got to know about the compute intensive nature of ML implementations
About this app, (django,react,tfjs) It takes a twitter handle as input and lists the most recent tweets along with buttons for each to predict toxicity using tensorflowjs's toxicity model (pretrained)
I plan to make a twitter bot with my learning so far.
Also this project initially had a much larger scope and I had to cut down/scrap most of it as i'm using only freely available services to host this and running ml on such services did not seem fair to me.