Let's talk

The conversations I want to have are with firms or in-house teams who are weighing whether to build a private AI assistant on their own content. Scoping a proof-of-concept, technical advisory on an in-flight build, second opinions on architecture — all welcome. Cold pitches for generic chatbot work are not.

How to reach me

• Email — [email protected]

I respond within two business days.

What to include in your message

• What you are trying to build, in one or two sentences.
• The domain and the type of content the system would sit on.
• A rough timeline — even "we don't know yet" is a useful answer.
• Whether you have an existing engineering team, and if so, what they own.

What an engagement looks like

A typical project runs six to ten weeks, organised into the phases described in the methodology: discovery, corpus analysis, architecture, build, evaluation, and handover. Pricing is project-based and scoped after a free thirty-minute discovery call. I do not bill hourly for delivery work; I bill against agreed deliverables.

For larger initiatives, I sometimes work alongside a client's existing engineering team in an architect-and-evaluator role rather than full build. We can talk through the right shape on the discovery call.

What I am not the right fit for

Generic customer-support chatbots, marketing assistants, and "plug GPT into our public website" projects. Builds where the client wants to ingest copyrighted third-party content they do not license. Sub-two-week scopes. If any of those describe what you need, there are firms in the region better suited — I will happily refer.