Worked on the frontend side of the project. This web app is a collaboration between me and a friend of mine. It is written in React using Typescript as the main language. The goal of the creation of this website is for us, the creators of this website, to showcase the implementation of Spotify Web API's features such as authorization and API consumption, and connecting it to Google Gemini's API. Some examples of API consumption include getting user's spotify data, user's top tracks and audio features, and user's top items and getting Google Gemini's AI chatbot capabilities to "roast" them. The data is provided by Spotify API through OAuth which was made possible by utilizing NextAuthJS. For the front-end, we use NextJS and for the server state handilng, we use TanStack Query (React Query). The website is able to generate an ID card for users based on their listening habits on Spotify. The web app takes the five track features from each track from the user's top tracks on Spotify, which include Happiness, Danceability, Speechiness, Accousticness, and Energy and these data are then inputted into a certain prompt that requests Google Gemini to "roast".

VexO, or Veritas Ex Oculi, is my very first machine learning project, and it was written in Python using Tensorflow. VexO started when one of my coworker asked if I could build a tool to differentiate AI-generated faces from real ones , I saw it as a unique opportunity to take on my first machine learning challenge. Although I had no prior experience in machine learning, I was determined to prove to him—and myself—that I could make it happen. This led me to dive headfirst into learning, watching numerous YouTube tutorials on machine learning fundamentals, and consulting various AI platforms like ChatGPT, Claude, Google Gemini, and Perplexity AI for guidance. Through a combination of intense research and iterative experimentation, I was able to create "VexO," a tool that successfully distinguished AI-generated faces from real ones. Completing this project not only demonstrated my problem-solving capabilities but also ignited my interest in machine learning, setting a strong foundation for future projects in this field.

I made this when I was bored. Couldn't find a good reason to come up with a good name for this program, so I left it with what it does instead as the name. Why? Because I think it's funny. This tool was created using Python and combining it with UB Mannheim's Teserract OCR to read texts that are present in an image, which are then sent to Google Gemini's API to be summarized. The images that are going to be processed are being taken from the user's clipboard, which is why it works best when paired with the Snipping Tool app that comes with Windows. The original purpose of this tool was to help my coworkers summarize a text easily, but since the text contains sensitive information, I ended up not recommending this tool to them and uploaded the code to my GitHub repository instead. I'm not sure if that's a good thing or a bad thing, but hey, at least this program is now publicly available on GitHub.