Portfolio

Some of my work can be found in my GitHub here – sorry if some repos aren’t documented!

Undergrad Capstone – Autonomous Drone

For my undergrad capstone, I worked with Brian Cheong, Gerry Chen and Sagnik Som to build an autonomous drone. We documented our progress halfway through the term with a video embedded below!

Custom Super Smash Bros. Melee Controller

I’ve been getting into semi-competitive Super Smash Bros. Melee recently, but the default GameCube controller hurts my hands to play on for more than a few minutes. So, spinning off some existing designs here, here, and here, I designed and built a custom, ergonomic controller! I finished putting it together hours before Canada’s biggest tournament, in which I won a grand total of 1 game. Yay! Some features include a 1.5mm galvanized steel switchplate with mechanical keyboard switches, a laser-cut acrylic face and birch plywood case, and a GameCube controller wire I desoldered from an actual controller.

Cross Control – Indie Video Game Project

In late 2021, I was a project manager and software developer in a team of around 7 software devs, artists, and musicians, working on a video game called Cross Control. The project was for a video game design course – as part of the course, we iterated through multiple playtesting sessions and presented to representatives from Zynga, Ubisoft, and some indie game studios. The devlog and playable builds are available at my friend’s itch.io page here! The trailer is below:

Bowling Robot

Three friends and I built a bowling robot that is controlled via Wii motion controls, again for a course project. Check it out!

Sensor Robot Pipes, at DGP

In early 2022, I worked with the Dynamic Graphics Project with Dr. Varun Perumal, supervised by Prof. Daniel Wigdor and Dr. Paul Dietz. We began designing a methodology to deploy multi-purpose sensor robots in custom, modular 3D-printed tubing. Deploying mobile sensor robots would allow sensing to be done wherever such tubes could be fixed, opening various applications such as monitoring soil in agriculture, collecting data for smart-home HVAC systems, or detecting deformations in a bridge. For this project, I was in charge of building software that would take in context-specific data to automatically generate 3D-printed parts, which would then assemble into the physical tube required to enable robots to travel and sense over a designer-defined space or path. I also made data visualizations. Some images are below! Dr. Perumal is away from Toronto at the time of writing, so we’ve been unable to complete our physical prototypes – we hope to wrap that up and publish sometime in 2023.

A visualization for LIDAR scan data!

Speech Accent Classifier via Machine Learning

In late 2020, three classmates and I produced a machine learning project capable of classifying accented speech. We used a convolutional recurrent neural network (CRNN) to process clipped mel-spectrogram images, and achieved solid results of around 80-95% holdout accuracy in binary classification tasks! Some images are below.

Our neural net architecture!
Our results, compared against a baseline SVM model!

Some Art!

I’m no classically trained artist, but my parents made me attend art classes as a kid and I still enjoy making drawings for fun, or to give to people as gifts. I also started making pixel art recently. Here are some examples!