About Me

I'm a Grade 12 student in Canada. My main interests lie in the applications of machine intelligence across different domains—finance, computer vision, and medical imaging. I'm fascinated by how the same underlying techniques can be adapted to solve vastly different problems.

Background

I started programming in middle school, initially just building small tools and websites. In high school, I came across YOLO and was immediately drawn in by what real-time object detection could do. That curiosity led me down the rabbit hole of convolutions, CNNs, and eventually deep learning more broadly. I found myself constantly asking two questions: how do these models actually work, and what real problems can they solve?

I work primarily with Python and C++. Most of my projects involve PyTorch for deep learning, though I've also spent time with web development (Next.js, TypeScript) when building user-facing applications.

Current Work

I'm currently a research intern at UTMIST (University of Toronto Machine Intelligence Student Team), Project Neura, where I contribute to an open-source medical image segmentation framework. The work has taught me a lot about how research code differs from production code, and how to design systems that other researchers can actually use.

Outside of research, I founded Amplimit in 2024—an effort to turn ideas from business competitions into reality. The first venture under Amplimit is LexAmp, an AI-powered legal services platform that provides legal consultation and smart contract generation. I'm also planning to expand into research tools and quantitative finance in the future.

Interests & Directions

On the technical side, I'm most interested in:

• Quantitative finance: Applying ML to market prediction, risk management, and portfolio optimization. Markets are noisy, adversarial, and constantly evolving—a challenging testbed for any model.
• Computer vision: I started with object detection, moved into medical image segmentation, and I'm currently exploring open-vocabulary image classification. I enjoy how CV problems are both visually intuitive and technically deep.
• Medical imaging: Particularly building tools and infrastructure that make research more accessible. There's something satisfying about helping other people do their work faster.

Get in Touch

I'm always happy to chat about machine intelligence, finance, or interesting problems in general. If you're working on something cool or just want to exchange ideas, feel free to reach out through i@stevenchen.site, or find me on: