About
I got into AI long before ChatGPT, working on traditional machine learning and deep learning.
When ChatGPT launched in November 2022, I tried it on the first day and realised this was something completely new, something that could change how the world works. So I dropped everything else and went all in. Before OpenAI released an official API, I built an unofficial one for a project I wanted to try.
Since then I've explored almost everything that has come out in the space: new models, frameworks, agentic coding tools. I even aced a semester final with a project I built that turned my professor's slide decks into podcasts, well before NotebookLM shipped the same idea.
What pulls me in is using AI to solve real problems and make people's everyday lives easier. So I keep building.
See what I've built in Work and where I've worked in Experience.
Shipping since 2022
- ChatGPT, day one
- Unofficial ChatGPT API
- Slides-to-podcast tool
- Published ML research
- More shipped in between
- Atlas, open source
- Geste, on the App Store
- Greenwash, open source
How I build
- 01
Use AI only where reasoning adds value. Keep everything else deterministic.
In Greenwash: the model handles judgment, code handles policy
- 02
Build systems that can verify, expose, and recover from failure.
In Atlas: a verifier that refuses, re-retrieves or escalates
- 03
Start with the problem, then choose the simplest architecture that reliably solves it.
In Atlas: one Postgres for text vectors, page vectors and the cost ledger
Credentials
Certifications
Recognition
- Amazon ML Summer School 2022Selected for Amazon's Machine Learning Summer School program.