Hi, I'm Gabriel St Pierre. I started my programming journey in the early 2010's, writing mods for minecraft. In 2023, I graduated from the university of Ottawa with an honors degree in computer science.

This website is used to gain a little bit of insight into my professional and personal life as well as a potential future blog spot.

Feel Free to contact me using any of the links above in the website's header.

Outside of programming, my main interest lie in sports (mainly formula1), creating house music, hiking and playing competitive video games.

My tech related interests evolve as I learn more and more about the field and emerging technologies. However right now, I'm interested mainly in functional programming, system design, data science / machine learning and full-stack development. Things I would like to explore in the future include embedded programming, advanced database design and video game development.

I'm always working on a project of some kind, many of which will never get finished. However below I've listed a few of my favorite projects that I've worked on in the past and finished. Below I've also listed the tech and tools most prominently used in each of the projects.

Handle sign detector

This project was used to gain an introduction into the world of computer vision by creating a program that detects different hand signs and runs specific functions for both. It utilized OpenCV alongside MediaPipe to detect the hand and the signs.

OpenCV
Python
MediaPipe

Live Streaming Site

This project was used as a way to learn about Azure Media Services, a cloud based media service that supports live streaming. It was built using AMS, Websockets for the real time chat as well as an express backend to handle authentication and data pertaining to users.

Websocket
Express
React
PostgreSQL
Typescript

Multilingual Emotion Detection (English/French)

This project set out to identify existing methods for creating Multilingual emotion detection models and to build a new one by applying similar methods to a relatively small dataset. Results were mixed, with the model performing on the low end of similar models that were trained on much larger datasets.

Python
Pytorch
Jupiter Notebook
Optuna