We developed NEWSFLASH AI, a news summarization app that leverages machine learning, Llama, and Pygame. The app allows users to choose their preferred news outlets and browse news in their category of interest. It then gathers news articles from those sources and provides summarized content with headers and bullet points for multiple articles. Leonor focused on the front-end development, creating the app interface, user screens, buttons, images, and text displays. This app saves users the hassle of reading articles across multiple platforms to obtain the same information, making it easier and more timely to stay updated with the news.

For this project, we created an AWS API that fetches the top news articles every day and stores them, along with their vector embeddings, which it uses to group articles together. After a user opens the app and specifies their preferred news sources, a call is made to the API that generates a headline with bullet points for each grouping (using only the specified news sources), which is then displayed in the app. The headers and bullet points are generated from two models: one that tokenizes and then combines semantically similar articles and one that feeds a Llama model the newly combined articles to generate headers/summaries.