This is a pathetic answer, but for me its probably an NLP web endpoint I built. The task was to take a query and categorize it into one of several categories provided when the server was started. So, for example, if a user submitted the query "barbie doll", the endpoint would return the three most likely categories: "toys", "clothes" and let's say "office supplies".

The way I did this was by using NLTK to compare the hypernym paths of the words in the query against the hypernym paths of the category names. I wrapped it in a tiny flask app and it was surprisingly fast enough for an MVP.