I embed with your customers and ship AI that works.
Hand me a fuzzy, high-stakes customer problem and I turn it into working AI software in your stack — fast. Agentic systems (autonomous agents, the harnesses and loops that run them, the evals that keep them honest), full-stack delivery, and the customer conversations that make it land. Twenty-five years shipping product, from the design file to the deploy.
AI-lab & startup FDE / solutions-engineering roles · in person across the US, or remote
The whole arc of an FDE, in one person.
A forward deployed engineer does three jobs at once: understand the customer, build the thing, and make it stick. I’ve spent 25 years doing all three.
Ships agentic AI end-to-end
Autonomous agents and the harnesses and loops that run them — with the evals and guardrails that keep them honest. Multi-step LLM pipelines, tool use, and retrieval, running in production against real latency and cost limits, not a demo in a notebook.
Customer-facing and technical
I sit with the customer, find where the real value is, and translate it into scope — then build it. I talk to engineers, designers, and leadership in each one’s own language, because I’ve held all three seats.
Fast in ambiguity
Fuzzy problem in, working software out. I scope tight, build in the open, and ship something real every few days — not a six-month roadmap that opens on delivery day and hopes for the best.
Full-stack, no handoffs
Front-end, back-end, serverless, data, auth, payments, deploy. I can own a customer solution from a blank repo to a live URL without waiting on anyone.
Design that gets used
People don’t use what they don’t understand. I make the surface clear and the onboarding obvious, so adoption actually happens after the demo ends.
Owns the outcome
As a founder I’ve answered for the P&L, not just the ticket. I care whether the thing works in the wild and moves a number — which is the only reason an FDE is in the room.
Things I’ve actually shipped.
Division 6 AI Visibility Audit
An end-to-end AI product I designed, built, deployed, and sell — a Claude-powered pipeline that audits how AI assistants recommend local businesses, delivered as an encrypted, paywalled report with Stripe checkout and a lightweight CRM. Read the technical case study →
Reffo.ai & Pelagora
Founded Pelagora, an open peer-to-peer commerce protocol, and built Reffo on top of it — an AI-first marketplace whose autonomous agent lists, negotiates, and closes deals inside seller-set limits. Architecture to launch.
P15 Studios · influence.INC · private roster
Games and products across a decade — Spell Royale (200K+ weekly players), Lexicube, and dozens of private builds for people and companies that will never have a public URL.
What I build with.
AI / agents
Engineering
Design & product
Range
I’ve sat in every seat at the table.
Most people selling AI have only ever sold. I’ve drawn the design file, written the code, and answered for the P&L — so I can walk into a customer’s world, understand it fast, and build the thing that fits it. Twenty-five years across film, games, and product taught me the one rule that matters here: people don’t use what they don’t understand. So I build things people use. That’s the job of a Forward Deployed Engineer, and it’s the work I want to do next.