HomeExperiencePapersOthersCV

Projects

sportBot
  • My lab mates and I were getting quite into badminton but it was difficult to book a court at the university gym, so I wrote a bot that would book courts for us. There are two parts to the bot, a scraper that will get the relevant booking information everyday and send it to slack, this script was ran using cron. Secondly, the booker that books the court chosen.

    • Capstone Project: Autonomous Vehicles
  • Vehicle tracking and occlusion handling are important issues in the field of self-driving cars. Occluded vehicles can result in the creation of new tracks. This reduces the efficacy of tracking algorithms. Together with a teammate, we create an algorithm to match tracks using heuristics based on the shapes and sizes of vehicles. We implored the use of the UnionFind algorithm to efficiently join tracks together.

    • Disentangling Sentences with Beta VAEs
  • Beta Variational Autoencoders (VAE) have been shown to be able to disentangle latent variable generative models, this means that individual latent traversals correspond to particular changes in features. For example, on the celebA dataset of faces, traversals can affect baldness or skin color, etc. We then used this framework on the sentence-VAE to try to disentangle sentences.

    • Gamecentre
  • An android app built with Gradle that contains 3 games. Minesweeper, Sliding Tiles, and Memory Puzzle.

    • Spotlight

  • Paired up with a partner company to create Spotlight, a social gym mobile app that allows individuals to check in at a particular gym and interact with one another. The app was created with a ReactJS frontend and a SQL backend. It was deployed to Heroku and Expo.

    • Guess the Drawing

  • Pictionary Style multiplayer online game where players send drawings of a theme to a friend and they will have to "guess the drawing". The web app was created with a ReactJS Frontend and a MongoDB backend. It was deployed on Heroku.


    • Volunteering

    Graduate Mentor, University of Toronto

  • I mentored a high school student over the summer of 2024 to investigate whether using RAG with chemistry datasets would give students better answers to high school chemistry questions. We coded everything from scratch and used pinecone as a vector database to store the dataset embeddings and gpt4o-mini as the base LLM.
  • He won second place at the competition!!

    • Curriculum Director, Project Include
  • I managed 15-20 university students every year to create coding curriculums for children aged 13 to 18 in Python, Java, and HTML/CSS. These university students would then present their classes at public libraries and priority neighborhoods. I got to teach a couple of workshops too.
  • My whole experience with Project Include was one of my highlights during university, the joy you get when a child understands lists and how they work is immeasurable.