Gordon Ng

Description

VSC (Valorant Store Tracker) was an open source iOS app that allows you to track your store and preview your skins from the game Valorant . It allowed for easy on the go access to your otherwise hard to access skin shop, a place to easily preview before buying, and an easy way to see all the content in the game. The application featured a modern UI design, implemented the MVVM architecture as well as async technologies.

VSC had a lifespan of a little under a year, amassing over 350,000 downloads and 50,000 active users. Towards the end of its lifespan, Riot Games and I came to a mutual agreement to take down the application due a conflict of interest with the game's official store. This application was a great learning experience and serves as the one of the cornerstones for my experience in iOS development.

Description

AI Game Agent is my deepest foray into AI, a project undertaken in the class COMP 424 Artificial Intelligence. It is an advanced AI developed for a strategic two-player board game. In this game, players move on a grid and place barriers to control space. The goal of the game is to wall off a maximum area of the board, with each tile counting as a point. This game agent achieved a 90th percentile score in a round robin tournament against other agents designed by my peers. AI Game Agent uses sophisticated game theory algorithms and heuristics to make intelligent decisions. These include:

• Alpha-beta pruning with a transposition table
• Iterative deepening search strategies
• Genetic algorithm with crossover and mutation to train heuristic weights

These methods enable efficient exploration of possible moves, optimizing decisions within limited time frames. Custom heuristic functions assess the game state from various strategic perspectives such as expansion potential, aggression, positional advantage, and barrier extension. These heuristics are weighted and combined to form a comprehensive evaluation of each move, making AI Game Agent a competitive and adaptable AI for grid-based board games. The genetic algorithm optimizes the heuristic weights of AI Game Agent. Key features include:

• A robust evolutionary approach fine-tuning decision-making against various enemy strategies
• Customizable parameters including population size, mutation and crossover rates, and number of weights
• Fitness evaluation based on performance against diverse enemy agents
• Evolutionary operations utilizing two-point crossover, Gaussian mutation, and tournament selection
• Diversity maintenance by periodically reintroducing successful individuals to the population
• Detailed logging to track best weights and win percentages against opponents for each generation

Description

CodeVault is a tool developed for the CSE challenge at the ConUHacks 2023 hackathon, featuring a React web app and an offline SwiftUI mobile app. It offers the latest password generation and verification tools, providing users with the ability to quickly create strong passwords and check their strength. Additionally, CodeVault offers valuable insights into cybersecurity, helping users make informed decisions about their digital security.

Our team embarked on this project to explore cybersecurity in software development. The web app was built using the React JS library, HTML, CSS, and JavaScript, while the mobile app was developed with Swift and SwiftUI. Despite facing challenges with React hooks and CSS design, the team successfully deployed both apps, achieving most of our initial goals. Through this project, we enhanced our React.js skills and improved our data parsing and analysis capabilities.