About

I build full-stack applications and applied ML systems. I don't fit neatly into one box — I work in Vue and PyTorch, design PostgreSQL schemas and train classification models, often on the same project.

My music taste is the same way. I'll go from nightcore pop to breakcore to power metal in the same afternoon. My code follows the same pattern — I follow whatever question pulls me next.

That's how I ended up building a semantic search engine. I started with topic modelling, ended up using an LLM to turn keyword clusters into readable labels, and then asked "how do I retrieve the paper's author?" — which became the search engine itself. One question leads to another.

Right now I'm thinking about optimization — specifically gradient descent as a heuristic method for niche problem spaces. I like divide-and-conquer approaches: quadtrees, recursive partitioning, anything that breaks a big problem into smaller ones until each piece is simple enough to solve.

I want to spend my early career building a broad foundation — across web, data, and infrastructure — and take ownership of problems end-to-end. I learn fastest when I can touch every layer.