I make AI systems businesses can actually rely on.

How I got here

I got into this early. At 16 I started teaching myself DevOps and a bit of programming, mostly because I wanted to understand how things actually run, not just how to build them. That turned out to be the useful half. Most people can get an AI to work on their laptop. Making it survive in production is a different skill, and that's the one I kept going deeper on.

At 18 I started an automation agency with a friend. In about six months we went from zero to real paying clients and over $20k in revenue, working with businesses in London and across Europe. Not prototypes or side projects — actual systems handling real traffic.

A few things I've built

• 01
  A content pipeline for a language teacher that pulled 1.3M views and 2,500 new followers in five days, with the top post hitting 865k views and a comment from one of the country's largest telecom brands.
• 02
  An automated video editing system for a London agency that took their turnaround from an hour per video down to about three minutes.
• 03
  Eval testing systems for Metra AI and its lead-automation product, so the AI's output could actually be measured and trusted instead of hoped for.

Each of these is written up in full, with the numbers, in the case studies.

What I actually do

What I actually do is the unglamorous part most people skip. Guardrails so an agent can't break your production. Human checkpoints where they matter. Eval tests that prove the output is correct instead of assuming it. And replacing expensive proprietary services with open-source pipelines, which usually cuts costs by a lot.

How I work

I work directly. No agency layers, no account managers, no telephone game. You talk to the person who builds the thing. That means straight answers about what AI can and can't do, fast iteration, and systems built to hold up under real use, not to look good in a pitch.