Frequently asked questions
What is a POC-to-Production Sprint?
A scoped engagement to get a stuck GenAI pilot into production: I find why it isn't shipping (it's rarely the model) and leave it running with spec-driven development, evals and guardrails.
What is a Fractional Head of AI?
I run the AI function on a fractional basis: roadmap, team and high-order decisions, without the full-time commitment of a permanent executive role.
Where do I start?
It depends where you are. If you don't yet know where AI creates value in your business, we start with an AI Discovery Sprint (find and validate the bet before building). If you already have a pilot that proved value and needs production, it's the POC-to-Production Sprint. If you need someone to run the AI function over time, Fractional Head of AI. Not sure? Email me and I'll point you the right way.
Who owns the code/IP and what happens after the Sprint?
The code and intellectual property are yours. When the Sprint closes I hand over what was built and documented for your team to maintain; I can stay on as Fractional Head of AI if you need it.
How do you handle confidentiality?
As standard: I'm happy to sign your NDA. Your code, your data and your IP are and remain yours, and I don't show your work without permission. If I share lessons, it's always anonymized.
How will I know it works?
Because it's measured, not promised. I don't hand over a demo that passed once: I hand over a system with its eval suite that checks whether it works, and I leave it observable in production. I build and measure, I don't just present.
How much does it cost and what is the scope?
Scope is bounded to a concrete goal before we start, and pricing is set by that scope, not by the hour: you know the cost before committing. Send me your case and we define it.
How long does it take?
It has a defined scope and a defined end from the start: focused and bounded work, not an open-ended contract that drags on. The timeline fits the scope we agree.
Do you work remotely?
Remote by default, and it's how I work best. On-site or hybrid is possible depending on the case and is agreed up front, along with the scope.
Glossary
- POC-to-Production Sprint
- A POC-to-Production Sprint is a scoped engagement that takes a stalled generative-AI pilot to production: it diagnoses why it is not shipping (rarely the model, more often evaluation, the surrounding context, cost, latency and data boundaries) and leaves it running measurably, with evals, guardrails and observability, and owned by your team.
- Fractional Head of AI
- A Fractional Head of AI is senior AI leadership on a part-time basis: it runs a company's AI function (strategy, roadmap, delivery standards and high-order decisions) without the commitment or cost of a full-time executive role, in the dose the company's stage needs.
- AI Discovery Sprint
- An AI Discovery Sprint is a time-boxed engagement to find where AI creates real value in a business and validate it is viable before building: it maps and prioritizes opportunities by value and feasibility, tests the chosen bet and ends in a production blueprint with architecture, risks, cost and plan.
- Production-ready in generative AI
- Production-ready in generative AI does not mean a system works once in a demo: it means it holds up in production. That is: evaluated with an eval suite that measures whether it works, guardrailed to bound its behavior against strange inputs and failures, observable when it degrades, within a cost and latency budget at real volume, and owned by a team that operates and evolves it.