AI product engineering practice

AI-as-a-Service: systems in production

Our AI work ends with a system running in production, a runbook and a written handover, not a slide deck. Week 1 ends with a plan and a price, in writing.

A paper engine house with vermilion gears, its handbook tied in a ribbon with a key on top, leaning on its door.

What the service covers

The order is the same every time: discover, design, build, evaluate, deploy, govern. What goes into each step is written for your business.

  • Architecture review

    We read what you run today and write down what to build, what to buy and what to leave alone, with the reasons.

  • A reference build in your stack

    Working code in your accounts, typed and tested, with logs and traces you can read from the first deploy.

  • An evaluation harness

    Test sets drawn from your own work, run on every change, with an alert when answers start to drift.

  • The handover package

    A runbook, an escalation path, the decision log and a model card, so your team can own the system without us.

3 ways to engage

Pick the one closest to where you are. Week 1 settles the scope, in writing.

  1. 1What you get

    Architecture review

    A written review of what you run and a recommendation you can act on, with or without us.

    How it runs

    One week of discovery that ends with a plan and a price, in writing.

  2. 2What you get

    Pilot to production

    A live system, built one slice at a time, with its evaluation harness, runbook and handover.

    How it runs

    Fixed scope and fixed cost, agreed before work begins, with a named person on every gate.

  3. 3What you get

    The embedded team

    A senior team on architecture, governance, vendor choices and escalations.

    How it runs

    Inside your organisation, in your accounts and on your direction. Documented handover, or we keep operating it with you.

Who the service is for

AI-as-a-Service / Fit check

It fits

  • Enterprises whose leadership has asked for AI in production, with a budget behind the ask.
  • Funded startups adding AI features to a product that already has customers.
  • IT services firms building AI for their own clients who want a second team on the design.

It doesn’t fit

  • A chatbot for your website. Other firms will build that for less.
  • An idea that has no product or customers yet. Our advisory practice is the better first step.
  • Pure research or paper-writing.

Questions about this service

Every engagement runs the same method: an agent proposes, a person approves.

How quickly can you start?
Discovery takes one week. We map the process, the data and the risk, and the week ends with a plan and a price, in writing. The build timeline is set there, against your team’s real readiness.
Which stacks do you work in?
Our own products run on Node and Python, with Postgres and Redis for data. In your estate we build in what you already run: the architecture decisions matter more than the language.
Can you work with our existing team?
That is the design. We pair with your engineers, ship with them and hand over the keys. We keep operating the system only if you ask us to.
Do you guarantee outcomes?
We commit to the agreed scope, not to business results that depend on your customers and your data. At the end you hold working software, an evaluation harness, a runbook and the decision log.
What does it cost?
Fixed scope, fixed cost, agreed before work begins. We don’t publish a rate card because engagements differ; the price is in writing in week 1.