01
Project
A complete AI system, end to end.
$5k – $50k+
Scoped before work begins
Chapter I
The Work
What I build, what it costs, and how to start.
II — How I WorkBuild it once.
Build it once.
Build it right.
No subscriptions, no vague hourly bills. I build a thing end to end and hand it over working — a chatbot that actually knows your business, a dashboard you can scan in a glance, an automation that retires a job nobody wanted to do.
Four ways to start, depending on how big the thing is.
A Good FitWho I work
Who I work
best with.
The engagements that go best fit a particular shape:
- —An established business with a working operation already in place
- —Years of accumulated data sitting in files, folders, or records that are hard to query
- —Budget for a real build (a local on-prem AI deployment runs ~$20k in hardware alone)
- —Looking for an upgrade to how the business runs, not a rescue from a crisis
If that's you, the call is worth having. If you're earlier than that — building from zero, no accumulated data, hunting for the cheapest path in — I'll say so honestly and point you somewhere better.
Three ways
to engage.
Transparent · Scoped before work beginsCurrently at capacityMy workload is full, so I’m not taking on new clients right now. Send an inquiry and I’ll get in touch when there’s room.
How long does a project usually take?
A typical Project runs 6–16 weeks depending on scope. The bigger builds — on-prem AI stacks, RAG systems with serious ingestion pipelines — sit at the longer end. I'll give you a real timeline with milestones before any money changes hands. Embedded engagements run monthly, indefinitely; the first month is usually onboarding-heavy and the cadence settles by month two.
Will you work inside our existing codebase?
Usually, yes. Most clients show up with something already running that just needs an AI layer or an automation added. I look at what's there, suggest the smallest change that solves the problem, and build it.
Which models and stack do you use?
I pick the model per task — Claude, GPT, Gemini, whichever fits. Stack is usually TypeScript, React, Node, Python, Postgres with pgvector, Supabase. For automation, n8n or custom code. Hosted on Vercel, Fly, or your own cloud.
How does pricing actually work?
Projects land inside a quoted range before work starts, and bill against milestones with a written scope. Embedded engagements are a flat monthly fee in the $8K–$15K band with a renewal cadence we agree on up front. Consulting is hourly with a 2-hour minimum. After a build, ongoing maintenance is on a monthly retainer with a written scope — predictable, not surprise pricing.
Will my data and IP stay private?
Yes. Happy to sign an NDA before we talk specifics. Your data stays in your infrastructure or in an isolated tenant — never mixed with anyone else's, never used to train anything public.
What happens after launch?
Every project includes 30–90 days of support (longer for the bigger tiers). After that, most clients move to a monthly maintenance retainer — written scope, predictable cost, no surprises. I don't disappear once the invoice clears.
V — Get In Touch
Got something to build?
A 30-minute call. No deck, no pitch — just tell me what you're trying to solve and I'll tell you if I'm the right person to help.
Send an Inquiry