Software engineering, AI & robotics
I build software, train models, and make robots move.
Mathematics & Computer Science at the University of Toronto.
A little context
I study Mathematics & Computer Science at the University of Toronto. Most of my learning starts with a question I can’t leave alone.
Right now, I’m building backend services at Fourth Dimension. In my research, I look at how models work—and where they get things wrong.
Selected explorations
Games, models, and a robot that dances. A few things I’ve built along the way.
Six mini-games, a fresh challenge every day, and a streak worth keeping. Built for iOS, Android, and web.
I connected daily challenges, streak tracking, and live Firestore rankings. Transactional score submissions keep repeated entries from spamming the leaderboard.
A market tool for Hypixel SkyBlock. Compare Bazaar prices with crafting and forging costs, then inspect profit, demand, and time requirements.
I built the interface and API service to bring prices and recipes together. Cached recipe data, fallbacks, and retries help handle unreliable upstream requests.
A computer-vision project that detects cars, people, and other objects in real time, with a live view of what the model sees.
I worked in a team of five to build and test the CNN model. We used real-time data collection and visualization to inspect its predictions.
A small robot with its own dance routine. An idea becomes something you can watch moving across a table.
I designed the 3D-printed body in Fusion 360, wired the electronics, and programmed the motors with Arduino.
Research & robotics
How a model performs across patient groups. How a robot responds to the world around it.
I built a GPU-accelerated CNN pipeline to distinguish melanoma from nevi using HAM10000, a source dataset of 10,015 dermatoscopic images.
The project evaluation reported approximately 0.90 AUC and 89–91% validation accuracy. A Fairlearn audit found a 4.45% accuracy gap and different false-negative rates by sex. The dataset total describes the source collection, not the size of a held-out evaluation set.
Designed, built, and programmed VEX robots and Arduino microcontrollers in C++, connecting code to motion on physical hardware.
I used sensor feedback and closed-loop motor control to make the robots respond in real time, with both autonomous and driver-controlled routines. Earlier VEX competition work qualified for provincials.
Beneath the surface
The tools I use, the teams I’ve worked with,
and a little about the person writing the code.
Technical foundations
The interesting problems rarely fit inside one discipline. These are the tools I bring together to solve them.
Six fields, connected through the work.I use these to build websites, mobile games, and the code that runs on a robot.
At 4D, I work on APIs, Docker services, RabbitMQ messaging, and database queries that keep requests moving.
I train and evaluate models, then look closely at their mistakes. My work includes object detection, skin-image classification, and fairness testing.
Tools I use for LLM applications, document retrieval, and AI-assisted development.
At QMorphix, I compared quantum and classical methods for reconstructing MRI images and reducing noise.
I build and program robots, connecting sensor readings to motor control so they can respond on their own or to a driver.
Experience
I’ve shipped backend services, tested quantum methods for medical imaging, and taught students to code. Here’s what I contributed in each role.
View résumé (PDF, opens in a new tab)Engineering · Research · Teaching
About me
I like understanding how something works, then trying to build it myself.
I'm studying Mathematics & Computer Science at the University of Toronto. My projects have taken me from a dancing robot to mobile games and medical-image research. I also enjoy helping other people get started with code, through teaching and volunteering.
Education · 2025–2029
Honours Bachelor of Science
Mathematics & Computer Science
Languages
Every ending is a trailhead
A team to join. An idea to build. A question worth following.
I’d love to hear what you’re thinking.
