Proof of work
What I've built
Five AI systems built and documented by me in the last eight months, plus two concepts. Two are live on public URLs you can open right now. Every entry is labelled for what it actually is instead of being dressed up, including the one whose study did not go my way.
Loading systems…
Why this is on the site
An advisor who has not shipped will cost you a year.
The reason most AI projects stall is not that somebody chose the wrong model. It is that nobody scoped the integration, the data was not ready, permissions were an afterthought, and no one planned for how the work would actually be adopted.
I have hit all of those problems personally, recently, and on my own money. That is what I am selling: the judgement that comes from having built these systems, not from having read about them.
A note on honesty, since it matters more than the portfolio. None of these are revenue-generating businesses. They are systems I built to prove a method and to solve problems I understood. Anyone telling you their side projects are thriving businesses is usually rounding up.
Recurring themes
- Multi-tenant architecture with strict data isolation
- Enterprise authentication, MFA, role-based access
- Real-time data across every connected client
- Third-party integration and workflow automation
- Cost-aware model routing instead of one premium default
- Billing and metering built in from the start