About

I'm a third-year Computer Science undergraduate at Monash University, and I've been here the whole way through. I'm about to start Year 3, Semester 2, with my Honours year coming next. My focus has narrowed steadily toward AI: less about using models, more about understanding why they work and where their structure comes from.

I've worked as an AI engineer in a mobile lab at RedNote (Xiaohongshu), one of China's largest social and lifestyle platforms with around 400 million daily active users, on on-device AI for mobile and portable devices. I built agentic systems, including GUI agents, that wired the company's internal tools together: a skill-management platform, MCP integrations, and other current techniques put into systems that had to actually ship.

A lot of why I care about the distance between a model that works in a paper and one that works in production comes from that stretch of work.

These days I'm most drawn to foundation models and multimodal learning. My final-year project is 2D-to-3D video conversion built on diffusion models, a genuinely new corner of the field.

Alongside it, I've been rebuilding reinforcement learning from first principles. The RL notes on this site are the result, and I'm carrying them forward toward PPO and beyond.

I build tools around my reading because I think better when ideas become inspectable. This site is part personal portal, part open research notebook. The notes are written to be rigorous enough that I'd trust them myself months later; the blog is where the thinking is looser and more personal.