Where technical AI
meets real business
problems
I build AI systems for enterprise. Not demos, not proofs of concept. Real products deployed in high-stakes industries where getting it wrong has consequences.
Who I am
I'm a founder and AI product builder. My work lives at the intersection of technical AI, organisational psychology, and enterprise strategy. I care about the gap between what AI can do in a lab and what it actually does inside a company with 50,000 employees, legacy systems, and real constraints.
I've co-founded and exited an AI company. I've built psychometric engines, clinical simulation platforms, and workforce intelligence systems from architecture through to enterprise deployment. I think across disciplines because the best solutions come from connecting fields, not siloing them.
Intelligence should compound
Every people decision should make the next one better.
AI is infrastructure, not spectacle
The value is in context and data flywheels, not the model.
Mistakes are data
Move fast, break things with intention, learn from every failure.
We'll figure it out
Comfort with the unknown isn't a personality trait. It's a practice.
Most AI projects struggle not because the technology doesn't work, but because it's disconnected from an actual business need. The work I care most about sits in that gap: understanding the technical architecture deeply enough to build it, and understanding the business well enough to explain why it matters. That translation work is unglamorous but it's where most of the value gets created or lost.
I don't wait for perfect conditions or complete information. I'd rather start building with 60% clarity and learn the other 40% by doing. Every mistake sharpens the direction. Every unknowable becomes knowable once you're inside it. The discomfort of not knowing is temporary. The cost of not starting is permanent.
Tennis, systems theory, organisational psychology, philosophy. I'm not an expert in any of them. But I've found that the most useful solutions come from importing a half-understood idea from one field into a problem in another. The willingness to be a beginner in multiple places at once turns out to be surprisingly productive.
Get to know me
Navigate through to see how I think and work.
Mistakes are my favourite teacher
I've shipped things that broke. Built features nobody used. Pitched to rooms that weren't buying. Every one of those taught me more than the successes did. I'd rather move and learn than wait and wonder.
I like building things myself
React, Python, system architecture, enterprise sales decks, pricing models, psychometric algorithms. I don't naturally hand things off. I'd rather build them, break them, learn from the breaks, and ship again.
Hyrox is my competitive outlet
Running plus functional fitness, under race conditions. Sled pushes, wall balls, farmer's carries. It demands the kind of conditioning that doesn't let you hide a weakness. Like building a company.
I see the Renaissance everywhere
The printing press, the rise of polymaths, the collapse of gatekeepers. AI is following the same script. I think about this more than is probably healthy.
The gap between technical and commercial
Most AI work dies in the gap between what engineers build and what businesses need. I've spent years learning to live in that gap. Designing psychometric models in the morning, presenting ROI cases in the afternoon. Neither skill matters without the other.
The unknown is where the interesting stuff lives
I've never regretted diving into something I didn't fully understand. Waiting for certainty is just another way of standing still. The willingness to be wrong is what makes it possible to eventually be right.
My best ideas come mid-workout
Something about a heavy set or a long rally on the tennis court unlocks a different kind of thinking. I keep voice notes going between sets. Most of my product architecture started as a breathless memo in the gym.
Not everything should be fun
I actually hate running. Genuinely. But I do it because Hyrox demands it and growth lives in the things you'd rather avoid. The gym, tennis, a heavy sled push? Love it. The running? Suffer through it. That's the point.
AI + Renaissance
The printing press didn't just make books cheaper. It democratised knowledge production, collapsed the distance between ideas and audiences, and made cross-domain synthesis possible for the first time. Polymaths didn't emerge because people got smarter. They emerged because the infrastructure for it finally existed.
AI is doing the same thing, faster. It's reshaping who can create, how quickly ideas compound, and what kind of thinking becomes possible when the cost of knowledge work approaches zero.
The printing press
Collapsed the cost of reproducing knowledge. Made ideas portable, shareable, and composable across disciplines.
Foundation models
Collapsing the cost of generating and synthesising knowledge. Making expertise composable and domain boundaries porous.
Rise of the polymath
Da Vinci, Alberti. Individuals who synthesised art, science, and engineering into forms nobody had imagined.
Rise of the generalist-builder
Individuals who move across product, design, engineering, and strategy because AI makes each accessible at higher bandwidth.
Disruption of gatekeepers
The Church's monopoly on knowledge distribution eroded. New institutions and new forms of authority emerged.
Disruption of expertise silos
Credentialed expertise is being unbundled. The moat shifts from knowing things to contextualising and applying them.
The outsider who built the impossible
The problem nobody could solve
Florence's cathedral had a 42-metre hole in its roof. The greatest architects in Europe agreed it was impossible to dome without the traditional wooden scaffolding, but no trees were tall enough. For 140 years, the cathedral stood open to the rain.
The experts couldn't see past their training
Every trained architect approached the problem with the same framework: build scaffolding, then build the dome on top of it. They couldn't solve it because they couldn't step outside the assumptions of their own discipline.
An outsider imported from another field
Brunelleschi was a goldsmith and clockmaker, not an architect. He invented a double-shell herringbone brick pattern where each layer supported the next, eliminating scaffolding entirely. He brought metalwork precision to a masonry problem.
The impossible, built
The dome still stands 600 years later as the largest masonry dome ever constructed. It wasn't built by the most credentialed person in the room. It was built by the one willing to import ideas from where nobody else was looking.
What I work with
builder
AI & Product
Engineering
Strategy & Commercial
Hyrox & the long game
I compete in Hyrox, a fitness race combining running with functional workout stations: sled push/pull, rowing, wall balls, farmer's carries. I'll be honest, I hate the running. But Hyrox demands it, and growth lives in the things you'd rather skip.
Outside of Hyrox I play tennis and train five to six days a week. Structured resistance work, simulation sessions, and the court when I can get on it. My best product ideas come mid-workout. Something about a heavy set or a long rally unlocks a different gear of thinking.
Showing up every day regardless of conditions is the same discipline that gets a company through the hard middle. The gym is where I process complexity. The court is where I test instinct.
Right now
Generative UIs will change every interaction we have online
I'm deep in research and hands-on building of generative interfaces. Not chatbots. Fully dynamic, context-aware UIs that reshape themselves around the user. I believe this will transform every layer of digital interaction, from pre-sales touchpoints long before a customer ever sees the product, through onboarding, through the product itself. Every experience becomes unique. That's the unlock most people haven't grasped yet.
Preparing for the next Hyrox
Back in race prep mode. Dialling in the running splits, building sled strength, and working on the transitions between stations where most people lose time.
Learning French
Adding a new language rewires how you frame problems. French forces a different precision of expression. Plus, there's something about committing to being terrible at something new that keeps the ego honest.
AI as experience architecture
How AI changes the value chain of products. Not just what the product does, but how it's discovered, evaluated, sold, and experienced. The interesting work is happening in the layers most builders ignore: the pre-sales experience, the first-touch personalisation, the compounding intelligence that makes every interaction different from the last.
Going deeper on the Renaissance
- The Civilisation of the Renaissance in Italy
- The Renaissance Bazaar
- Brunelleschi's Dome: How a Renaissance Genius Reinvented Architecture
Not casual reading. Primary sources and deep history to understand the structural mechanics of how an entire civilisation reorganised itself around new knowledge infrastructure. The parallels to AI keep getting sharper.
Ideas I keep returning to
Technology without business context is just a demo
Most AI projects struggle because the capability is disconnected from the actual enterprise need. The translation layer between what AI can do and what a business needs it to do is unglamorous work, but it's where most of the value gets created.
Knowledge cost approaching zero
When generating knowledge becomes nearly free, the scarce resource shifts to judgment, taste, and asking the right question.
We'll figure it out
The most powerful sentence in any room. Not reckless optimism. Earned confidence that any problem is solvable if you stay with it long enough. Applies to products, relationships, life decisions, all of it.
The data flywheel as the real moat
Models are commoditising. The defensible advantage is organisational context that compounds. Systems that get smarter with every decision they support. The winners won't have better models. They'll have better data loops.
Feedback loops over features
The founders who think in leverage points outlast those who think in feature lists. A company is a system of systems.
Mistakes are the curriculum
I've shipped things that broke, pitched to rooms that weren't buying, built features nobody used. Every one taught me more than the wins did.
Conversations
It's never been more important to be a beginner
I have no experience in hardware. Zero. And right now I'm spending my evenings trying to connect LLMs to physical devices. I'm bad at it. I'm learning in public. I'm making embarrassing mistakes. And I think that's exactly the point.
Can I develop this knowledge with the AI tools available to me? Am I shameless enough to try? That's the new question. Not "do I already know this?"
I'd rather be the person who tried ten things and failed at six than the person who perfected one thing and missed the other nine. The tools to figure things out exist now. The only thing stopping most people is the willingness to look stupid for a while.
Let's talk
AI, enterprise strategy, building in the unknown, or something I haven't considered yet. Good conversations start with curiosity.