Work
Payments is the domain; AI with judgment is the method; APIs, webhooks and MCP servers are the surface where both meet. Each case below is a receipt for the same thesis — the decision that moves money stays deterministic and auditable, the model lives at the edges, and evals decide when something is done. Read the routing case first: it is the spine of the portfolio, and its research version is under way as my MSc thesis.
Method, not product: how I work with AI.
- Payment routing an agent can operate — evidence in the hot path, AI at the edges
A routing engine that decides which payment service provider should receive a card authorization, and what happens next if it declines — on empirical approval evidence, at a fee tolerance the operator states in percentage points. The thesis: the decision that moves money stays deterministic and auditable, and the language model lives only at the edges.
- nutri. — running a production app inside an agentic stack
A nutrition-tracking PWA I design, build and operate entirely inside Claude Code. The interesting part isn't the app — it's the operating system that keeps an AI agent effective and safe on a real codebase with no second reviewer.