Anthony Sa@Millionarc

I made an application that takes in datasets from a csv file 'input.csv' and displays visual aids of the data, allows the users to talk to their data through a llama chat bot that has access to the csv file. Creates AI recommendations for addons to their data, and which machine learning model they should use for accurate predictions. The AI also can infer and draw conclusions on the datasets to allow users to learn more about the data and what it really represents. Lastly, it writes Jupyter Notebook code for machine learning and AI training. It writes the code depending on the different columns in the dataset and bases the whole model around what it thinks you should be going for your output.

The Universal code obfuscator is designed to help both starter and intermediate coders with code obfuscation, especially when ran in the browser. It works by prompting Codestral with the file and then trimming the response given back. I do want to let the judges know, for some reason since last night Codestral wasn't returning code on majority of my requests (1/20), so I changed the font to something less specific and less useful for now. I contacted Codestral support and I'll let a mentor know if they reply so they can alert the judges. If executing the demo on your computer, feel free to try "prompt1" as this one is extremely more effective than the other.

My idea is to have a medical assistant web application where anyone regardless of computer experience will be able to figure out what medical issues they have and what steps they should take to help their symptoms/condition. This is meant to help lower income Americans who are unable to go to the doctors office for minor medical issues which could be major. The flow of the website and UX is all done in LangFlow as seen on the presentation slide and an exported JSON in the GitHub repo. The webapp takes in the different symptoms of the user, the severity and duration of the symptoms, additional info about themselves/the issue they are having, and an optional image upload.