Most career summaries are unusable. Mine included: "product vision", "cross-functional leadership", "0→1→scale". True, and interchangeable with ten thousand other people's.
So here are five actual decisions instead. Each one is a moment where the obvious choice was available and I took a different one. They're the stories I'd tell you in a room, written down.
The four-month clock: how a hardware programme reached manufacturing readiness in four months instead of eighteen
Most hardware programmes budget eighteen months to get from working prototype to manufacturing readiness. At Elsight we did it in four.
Not by cutting corners. By compressing the feedback loop between prototype, QA and the production line until those three stopped being sequential phases and became one conversation.
The standard model runs them in series: engineering finishes, hands to QA, QA finishes, hands to manufacturing. Every handoff is a queue, and every queue is where the eighteen months actually goes — not in the work, in the waiting. We ran them concurrently, which meant accepting that some prototype work would be thrown away because manufacturing feedback arrived before the design was final. That waste was cheaper than the queue.
Three hundred units validated. Certification for runs of ten thousand and up. A commercial stream over $1M opened.
The transferable part: when a timeline looks immovable, measure how much of it is work and how much is waiting. It is almost always mostly waiting, and waiting is a process choice, not a physical constraint.
The boardroom shift: the single slide that changed an investor conversation
The investor meeting at Ottopia was tense. Reasonable questions, sceptical tone, the specific kind of tension where everyone is being polite about doubt.
Then we showed the holistic-approach diagram — one picture of how remote assistance for autonomous vehicles fits together end to end, across three products rather than one.
The conversation changed in the room. It went from "is this technically possible" to "how fast can you scale this". Same company, same technology, same quarter. What changed was that the investors could now see the shape of the thing.
According to the company's CBO, that vision was credited with doubling Ottopia's valuation in investor conversations.
The transferable part: investors and boards are not evaluating your technology. They are evaluating whether you have a coherent model of your own market. A diagram that makes the model legible does more than a demo that proves the technology works, because the technology working was assumed and the model was not.
Show, don't tell: the forty-slide deck I threw away
At SheeldS we sold cybersecurity to heavy-commercial fleet operators. The deck was forty slides. It was accurate, thorough, and it converted almost nobody.
So I scrapped it and replaced it with a live hack simulation. We compromised the prospect's own fleet, in front of them, in real time.
Watching your own trucks get taken over is not an argument. It's an experience, and it doesn't need a slide to land. That change closed more deals in a month than the deck had in a year, and it fed a roadshow that came back with over twenty qualified enterprise leads.
The transferable part: a deck explains a problem; a demonstration makes the buyer feel it. If your product's value depends on the buyer believing a risk is real, stop describing the risk. Cybersecurity is the obvious case, but the principle holds anywhere the pain is invisible until you point at it.
Cultural product fit: rewriting a pitch on the flight to Tokyo
We took a Harman product on a roadshow through the US and Japan with one narrative: speed. Fastest to integrate, fastest to deploy, breakneck pace.
That worked in the US. On the flight to Tokyo I realised it was going to fail, because the Japanese automotive market was not buying speed. It was buying proven reliability — and a pitch built on "we move fast" reads, in that room, as "we have not finished testing".
So I rewrote it in the air. Same product, same slides mostly, entirely different emphasis: maturity, validation, what had already been through the wringer and survived.
That rewrite is what secured the first major OEM pilot. The platform went from one automotive OEM client to fifteen, scaling past 100,000 vehicles, with 1400% revenue growth.
The transferable part: localisation is usually treated as a translation problem. It's a positioning problem. The same feature is a strength or a red flag depending on what the market has learned to fear.
The deep tech advantage: why starting in firmware still pays off
I started in firmware. Twenty-seven years later it's still the thing that most differentiates how I work.
It means I can argue latency with an engineer and ROI with a CFO in the same hour, and be taken seriously in both conversations. Engineers stop performing simplification for me. Executives get a technical claim translated into a business consequence without a layer of hedging in between.
The specific value is not knowing the details. It's knowing which technical constraints are real physics and which are current implementation choices — because the second kind can be negotiated and the first kind cannot, and most product disagreements are really an argument about which one you're facing.
The transferable part: if you came up through engineering, don't let it fade into "technical background" on your CV. It is the thing that lets you tell a genuine constraint from a habit, and that judgement gets more valuable the further you get from writing code.
These five are also hidden on my CV as hover-activated easter eggs. Find all five there and something happens.