In 2017-2018 I took the Springboard Data Science Career Track course. I was working business news company, Curation.
At the time. Curation was paying a sizable team of journalists to select, summarise and tag news articles, an operation that was costly and non-scalable. I built this semantic classifier using machine learning techniques to demonstrate how the company might automate much of its editorial process, and reserve human intervention for editorial sign-off and original business insight instead of using it to perform relatively low-level text processing tasks. Several components are required in order to solve this problem.

Github links:
News Classfier Part 1
News Classifier Part 2