I graduated with a Bachelor of Science degree in Computer Science from the University of California, Los Angeles (UCLA) in August 2023. Prior to UCLA, I transferred from Long Beach City College (LBCC), where I earned two Associated of Science degrees in Computer Science and Physical Science in 2021.

Courseworks taken

BruinNotes is a platform for UCLA students to share, archive, and access course notes. It offers a user-friendly interface for adding courses, sharing and requesting notes, as well as engaging with content through likes, dislike and comments. The project was developed under Professor Miryung Kim’s guidance in CS 130.

Dodge, is a computer graphics game created with WebGL where players try to stay alive as long as possible by dodging the blocks and accumulating more scores. The project is supervised by Professor Asish Law in Computer Graphics (CS 174A) course.

Wordle Plus is an implementation upon the original "WORDLE" app with additional features that allow players to play the game with either four, five, or six letters WORDLE and as many times per day as they would like. The project was supervised by Dr. Paul Eggert for the Software Construction (CS 35L) course.

For this project, I enhanced a TI-RSLK car by incorporating phototransistors as sensors to detect variations in light reflection from a custom line track. Using Arduino software, I then reprogrammed the car to enable autonomous navigation along the track. This was part of the ECE3 course, supervised by Dr. Dennise M. Briggs.

Supervised by Professor Peter Reiher, our team uncovered four major security problems of curl 8.0.1 through web searches, a code review using various automated tools to find potential undocumented flaws, followed by live testing in a virtual machine to produce a demonstration of the vulnerabilities. From our investigation, we found four main vulnerabilities and some security concerns in the automated tools’ reports. Overall, we concluded that curl is a fairly secure system, as the vulnerabilities are not too severe and the code does a sufficient job of handling edge cases and preventing unexpected failures. Generally, any vulnerabilities that we did find were low impact and/or required very special conditions to exploit.

Our project aims to classify U.S. state names on license plates using machine learning algorithms. We’ve built and compared three models: Convolution Neural Networks (CNNs), Feed Forward Neural Networks (FFNNs), and pre-trained CNNs - ResNet50. These models were evaluated based on F1 scores and accuracy, and tested on a diverse dataset of license plate images. The best performing model was ResNet50 with a test accuracy of 85%.

I'm not a professional artist but I do love to paint during my free time so, here are some of my artworks.