AI products · built and working today
I design and build AI systems for banks, governments, and companies where mistakes are expensive. Everything is tested like software, checked by people, and explained in plain language.
Work
All six are built and running — and they connect like one system. Every one is explained below in plain language — no jargon required.
Country reports you can trust. Compass writes an economic briefing in seconds using live official data — and a built-in fact-checker deletes any sentence it can't prove. Tested on real data for Mexico, Colombia, and Brazil.
Read more →Never miss a new regulation. Radar reads Latin America's official government bulletins every day and sends a short alert: who is affected, what they must do, by when.
Read more →Quality control for AI. Every update takes an exam before it ships — if the score drops, the update is blocked. No more "it seems fine."
Read more →Proof that your AI follows the rules — a live inventory of every AI system a company runs, sorted by risk under the EU's AI law, with the evidence regulators ask for.
Read more →Fairer loans for small businesses. AI reads the paperwork in minutes and an interpretable model scores it — but a person always makes the final decision.
Read more →Practice big decisions before making them — a simulation where realistic regulators, rivals, and markets react to your plan, turn by turn.
Read more →How I work
Every product takes the same exam on every update. Scores are published — including the failures.
When the AI isn't sure, it says so and hands the case to a human. Uncertain answers never ship silently.
Every product page lists what it can't do yet. Trust is built on what you admit, not what you promise.
About
I'm a business administrator turned AI builder. From 2024 to 2026 I ran Enpath, a market-intelligence venture helping Korean companies weigh entering Latin American markets — research, partner due diligence, and AI-assisted workflows in Python and SQL. It didn't survive as a business, but running it end to end taught me where analysts lose their time and what AI genuinely fixes.
This portfolio is Enpath's direct descendant: Radar and Compass productize what I once did by hand, and Checkpoint exists because I never want to say "trust me" about an AI system's quality again. I'm drawn to fintech, public policy, and Latin America — and looking for my next chapter in AI product work. What I bring: curiosity, honesty about what I don't know yet, and working proof that I ship.
Have an idea, a role, or just curiosity about any of these products? Write me — I read everything.