Remote · Available for work
I build software that actually ships.
I design and build web applications, from marketing sites teams can run themselves to AI products running live in production.
What I do
Three kinds of work, one set of habits
Ship the whole thing, verify it actually works, and leave it maintainable by whoever comes next.
AI engineering
Retrieval systems, model routing, and agents that operate real software rather than just describing it. Grounded in sources, not model memory.
Full-stack web
Next.js and TypeScript, from the database to the interface. Auth, payments, and the unglamorous parts that decide whether a product survives contact with users.
Sites teams can run
Marketing sites wired to a CMS so the client edits their own content. No developer retainer for a copy change.
Selected work
Things I have built and shipped
A live AI product in open beta, a client site the client runs themselves, and the infrastructure underneath both.
Morgan
An AI tutor that operates a real ServiceNow instance alongside you
A tutoring product for ServiceNow certification, built around an AI persona that connects to the learner's own developer instance and works through tasks with them. Live, with real authentication (opt-in MFA, a guest trial that converts to a full account, proper account deletion) and a retrieval corpus behind every answer.
Silent Forge Technologies
A marketing site the client's team can run without a developer
A production marketing site for an engineering company with eight distinct product lines. Built so a non-technical team can add and edit everything themselves through a CMS, with no developer in the loop.
SimNow
A compatible sandbox, so learners stop waiting on real instances
A training sandbox that implements the exact HTTP surface Morgan's AI tutor uses to drive a ServiceNow instance, so the tutor runs against it unmodified, just by pointing at a different host.
Docs Knowledge Graph
GraphRAG over technical documentation: semantics plus structure
A retrieval layer that turns a large technical documentation set into a property graph alongside a vector index, so retrieval can follow structure (prerequisites, versions, cross-references) and not just semantic similarity.
Got something that needs building?
Whether it is a website, a web app, or an AI feature bolted onto something you already run: tell me what you are trying to do and I will tell you honestly whether I am the right person for it.