The engineer behind the work

Four decades of hard problems

I began with a five-year engineering apprenticeship at British Aerospace — materials, machine shop, avionics, automated test — before ever specialising in software, then worked on Airbus A310 and A320 fuel gauging, including three years at Airbus in Toulouse producing evidence for CAA certification. That grounding is why I design products, not only the software inside them — and why I prove the uncertain parts on the bench and let the result confirm what I already expect.

Since then the work has spanned aerospace, defence, telecommunications and industrial automation across the UK, Germany, France, Finland and Sweden, and more recently medical devices, mining, retail systems and hydraulic instrumentation — as an international engineering consultant, technical lead and systems architect. The industries have almost nothing in common; the method transfers.

How I work

My engagements are rarely a tidy specification handed over to build. More often a development has stalled, the original approach can't reach its objective, or the problem isn't yet understood. The approach is consistent:

• Understand the actual problem, not the stated one.
• Build a model of it that holds up, and separate its independent responsibilities.
• Generalise as far as the problem naturally goes — and no further.
• Where a design carries real uncertainty, build that part first, to prove it.

I don't arrive as the domain expert. I come up to speed in an unfamiliar domain fast, because the client's world is an engineering problem in its own right. And I use AI throughout — research, exploring alternatives, analysis, implementation — which has changed the rate at which a model can be refined, but not who is responsible for the judgement.

When to bring me in

• Product architecture from concept — deciding what should be built, and how, before commitment.
• Recovery of stalled or undocumented developments — reconstructing what a system is and does, then continuing it.
• Performance where the model is the limit — problems that won't yield to optimising the implementation.
• Technical due diligence — an independent read on whether a plan reaches its objective.
• Architectural direction alongside your own engineers, including writing the specification where none exists.

Geotech Design works remotely, worldwide, through Geotech Design Ltd (registered in England & Wales, No. 16436596).