From design to systems to AI — building things that hold up.

I started my career as a UX designer with a bit of application development on the side. After watching too many developers make excuses for incorrect and unoptimised implementations of my designs, I decided to go all-in on development.

Over time my interest shifted from building apps to architecting and implementing system designs, infrastructure, complex distributed systems and classical ML — NLP, CNNs and the like. Then the AI revolution arrived and, with LLMs becoming generalised, I went deeper into that space.

In January I trained my own SOTA Fusion-in-Decoder model — planning to open source it soon. Previously I built personal models for sentiment analysis, emotion classification and zero-shot classification. In the industry I've trained and shipped custom medical-grade classifiers, RAG systems and advanced automations for employers and clients.