I build software that turns scattered data and business rules into tools people can understand and act on. Applied AI is a specialization: I use it within workflows that keep decisions inspectable and people in control.
A job-search workflow with explainable matching and AI proposals that change nothing until approved.
An IT portfolio workbook rebuilt as a snapshot-based web application with role-checked editing.
AI can work with language. The harder engineering problem is making a request mean the right thing inside a particular domain.
Magic makes that problem concrete. A card’s meaning depends on its rules, timing, game state and interactions with other cards. A semantic ontology defines a domain’s entities, concepts and relationships. I’m exploring how to use one to connect card language to facts, constraints and executable behavior.
Owner-only hosted tool · recorded public builds
Grounded AI reasoning and deterministic constraints turn a strategy into a reviewable deck.
Single-host pilot · recorded historical runs
Turning approved bug reports into independently verified draft pull requests.
Where the research comes together
Deck Lab is the application layer: card research, deck editing and supported analysis in one workflow. The direction is to bring domain facts, semantic interpretation and decision tools together in a product people can use.
A competitive Commander research and deck-editing workbench built around explicit data and service boundaries.