About me
This site has stuff on it that I think is cool. Because it is cool.
A BBC Micro at an early age, then mechatronics, then motorsport electronics, then Formula One, then thirteen years building a software company and selling it, then an MSc in AI, and now robots. Throughout: a workshop, and a persistent inability to leave things intact.
The through-line
These jobs are in quite different industries, but the problems have often been surprisingly similar. Something has to work in the real world, and you need to find out whether it actually does. Engine test rigs, payment platforms, robots - different versions of much the same problem. How do we know this is really working?
1984-ish
My dad bought me a BBC B micro, I loved reading the manuals - they included details about the hardware and how the machine worked. I typed in listings from magazines and books, and really enjoyed playing with "my computer" - now there seems to be a movement that says "this isn't your machine" - my hardware, I can do what I want with it.
School
I made a model hovercraft for GCSE - including building my own PWM speed controller. And a robot CD player at A level - an arm would pick up the CD, place it on a player, lower the puck and then play it (controlled by a BBC B).
1997
BEng in Mechatronics: I was instantly drawn to this degree, it suited me perfectly, and it was great meeting other people who also knew the Maplin catalogue inside out. My final-year project was a robotic chessboard, which won an award and moved the pieces itself.
1997–2003
Pi Research (later Cosworth). Embedded C, engine controllers, sensors, wiring looms, CAN and serial integration, telemetry, fuel simulation, engine-test automation.
Trackside debugging is quite stressful and high pressure. You need to keep a clear head and arrive at a solution in a short period of time.
2003–2006
Two years at Jaguar Racing programme-managing the cross-functional design, manufacture and build of race-ready cars, then R&D at Red Bull Racing on sensors, actuators, software, hardware and automated systems.
F1 is interesting: the deadlines don't move, everybody is very clever and motivated, and you get tested every two weeks in a way that isn't arguable.
2006–2019
Thirteen years co-founding and building a secure B2B SaaS company, as CTO and chief architect, and hands-on the whole way. We ended up with 40 staff, £8M ARR, and a patented PCI voice-payment platform carrying more than half of UK utilities' phone payments.
I learned that a good company comes down to hiring good people and trusting your teammates.
2019–2024
Aeriandi merged with Voxygen to become Speik, which was then acquired by Dubber, where I ran technical strategy, platform reliability and compliance for EMEA. Around 300 production services across AWS, Azure and on-prem.
Also, in parallel, I was principal contractor on a self-built family home. The one with the curved walls, which definitely added to the complexity.
2025–2026
An AI residency at the ML Institute, then an MSc in Artificial Intelligence at Leeds, finished with Distinction. Nine taught modules and a dissertation on robot learning: MuJoCo simulation, teleoperation through a VR headset, PyTorch policies, and then the awkward business of testing them on hardware that has friction in it.
Now
Staff engineer at NeoIntralogistics, building warehouse systems that coordinate autonomous mobile robots and Universal Robots arms.
The rest of the time I am in the workshop, and that is what most of this site is about.
What I care about
I want to know whether a system is actually better, not whether the graph looks nicer.
Trust, challenge, humour and people playing their position on the pitch.
Software is more interesting when it moves hardware, handles real users or has to survive reality.
Contact
Oxford, UK. Happy to talk about robots, reliability, steel, or why your simulator is lying to you.