A decentralized Web3 crowdfunding platform built with Solidity and Flutter, featuring on-chain governance, milestone-based fund release, and ERC20 token issuance per project.

A research project analyzing gender bias in the COMET-ATOMIC20 commonsense knowledge model using Python. Explores differences in generated inferences based on gendered inputs using sentiment analysis and lexical methods.

The primary objective is to develop a robust database system to manage a voting process for a performance show featuring singers and bands. The system was designed to handle complex relationships and provide real-time voting results, ensuring a transparent and efficient process for judges and audiences alike.

As part of my work with Minivillage, a social purpose tech startup, I conducted a comprehensive analysis of their beginning-of-pilot survey.

This project presents a deep learning method to enhance compressed images specifically for machine learning tasks, addressing the performance degradation caused by image distortions such as blurring and blocking. The approach combines an image restoration network and a classification network to optimize compressed image quality for machine consumption.

This project aims to protect digital privacy by using a deep learning model, YOLOv8n, to detect sensitive information in WeChat screenshots. The model addresses privacy concerns by automatically identifying and allowing anonymization of personal details, enhancing security in digital media sharing.

CineMemo is a full-stack web application designed to provide users with a seamless experience for discovering movies, adding films to a personal watch list, and logging reviews for watched movies. The app is built with a responsive and dynamic front-end using React, while the back-end is developed with Spring Boot, allowing for efficient data management and integration with third-party movie APIs.

This project was a hands-on exploration of Artificial Intelligence, completed as part of CPEN 502 - Architecture for Learning Systems. The primary objective was to develop an intelligent tank, leveraging Reinforcement Learning (RL) and Neural Networks (NNs), to autonomously battle against explicitly programmed tanks within a game environment. Through this project, I applied machine learning techniques to design an unsupervised tank that can adapt to its environment and optimize its performance over time.

This project presents a deep learning method to enhance compressed images specifically for machine learning tasks, addressing the performance degradation caused by image distortions such as blurring and blocking. The approach combines an image restoration network and a classification network to optimize compressed image quality for machine consumption.

A full stack web application to help users easily access and analyze information about UBC courses and room availability. Using jQuery and TypeScript, I built a responsive and user-friendly interface, allowing users to navigate the data with ease. On the back end, I implemented the application with Node.js to handle data processing and respond efficiently to user queries.