Leave the world better than you found it.

Four years of mechanical engineering, and the parts that taught me most happened outside the syllabus. I led the robowars team — designing and building combat robots that had to survive an arena, on a student budget, against other teams doing exactly the same thing. It's an unusually honest form of engineering. Your design either holds when something hits it or it doesn't, and no amount of explaining changes the result.

I also co-founded the Entrepreneurship Development Club, which in hindsight was the first strategy work I ever did: taking twelve members' rough ideas and helping structure them into applications strong enough to win government funding.

My first job, and the one where I was bad at my job for the longest. Six weeks in, most of my peers were well past three hundred thousand in revenue and I was under a hundred. I was following the playbook precisely, and it wasn't working.

What changed it was asking for help. I brought an experienced colleague along to a demonstration and watched him run it, and the gap became obvious: I had been explaining a product when I should have been understanding a household. We closed a high-value deal in that session. After that the numbers came — and later I was able to pass the same shortcuts to a teammate on a performance plan. He made manager eight months after I left.

I also moved from direct field sales into inside sales, which is a genuinely different craft. You lose the room, the body language and the kitchen table, and you have to build the same trust down a phone line.

I co-founded a cloud kitchen around two things that shouldn't be in tension: employment for women and food people could actually eat. Fourteen cooks — we called them supermoms — a fifteen-person team, twelve products, on ten thousand dollars raised from friends and family.

It was as much a technology business as a food one, and I owned that side. We built the mobile app and the web storefront from nothing, ran the ordering flow, the catalogue and the social channels, and used what the data told us to decide what to cook next. One campaign was reshared by the state's Chief Minister and pulled in over fifteen thousand organic interactions.

We chose the hardest category deliberately. The logic was that if we could solve quality control in food, every other home-made category would be easy afterwards. That reasoning was backwards. Food has almost no shelf life, home cooking varies day to day with volume, and a cancelled order is simply a loss. The thing that closed us was the thing we'd assumed we would solve along the way.

Years of losses and a management team seriously discussing closure. Before running a single sales session, I sat with the sales team and asked them why they thought we were losing money. They already knew. Nobody had asked them.

The answer was the customer. We were selling one laptop at a time to walk-in consumers in a market where enterprises and government departments bought the same products in volume, at better margin, on repeat cycles. I repositioned the business toward B2B and B2G and rebuilt the sales process, client engagement model, CRM, website and supplier terms around that decision.

Then a second opportunity surfaced. Server demand was climbing ahead of the World Cup and the margins were far better than anything we carried — but we had no technical capability and no spare capacity to build it. I taught myself server hardware on YouTube, started quoting RFQs myself, closed a deal with advance payment, and brought in freelance technicians to install.

I had been running businesses on pattern recognition and nerve. The MBA gave me the machinery — discounted cash flow, market sizing, structured problem solving — and, more usefully, made me defend my reasoning in front of people who were better at it than I was.

I won two case competitions, which is the closest thing school offers to the real job: a messy brief, a hard deadline, and a panel who will find the weak assumption if you leave one in. I finished near the top of the cohort and taught three of the courses back as a teaching assistant — marketing, entrepreneurship and data analytics. Explaining regression to twenty people who haven't seen it before is its own kind of stress test.

Five-plus external mandates in fifteen months, run alongside the MBA and a full-time analyst role. Go-to-market, product launch, regulated market entry — real clients, real boards, decisions that happened afterwards.

The mandate that taught me most was a cannabis dispensary licence in a jurisdiction that hadn't finished writing its cannabis regulations. Nobody on the team wanted to lead it, because there was nothing to research. I took it, and made a habit of calling the authorities who were actively drafting the rules whenever we hit a grey area. It gave us direction; more than once it showed them a gap they hadn't accounted for.

The other lesson came from a $25M European garment steamer manufacturer entering North America. The data said the consumer market was crowded. The useful answer was somewhere else entirely.

The Defence business held a market analysis commissioned from a top-tier consulting firm. It was rigorous, and it was already going stale, because it was a document. Refreshing it meant commissioning it again.

I rebuilt the logic as a live Power BI model running on Cirium fleet data, so the opportunity set recomputed itself as the fleet moved. The part I'm proudest of wasn't the dashboard: it was a method for estimating aircraft retirement age by country, which turned a snapshot of what exists into a forward view of who has to replace what, and when. That's where the untapped opportunities were sitting. I presented it to the Director of Defence Programs.

I worked on the Aftermarket side of the business as well, hosting the daily S&OP meetings across operations — aligning teams who each held a different, entirely reasonable view of the same week. That's where I learned that alignment is a harder problem than analysis.

Engines fly for decades, and the interesting economics happen late. My work sits on mature and end-of-life programmes: which aftermarket commercial offerings still earn their place, what it's genuinely worth to keep a fleet flying, and what the numbers look like across a ten-year horizon rather than a quarter.

The horizons run longer than that when they need to. I've built models fifty years out — the timescale you need when the conversation is a partner negotiation and the contract at the end of it is measured in billions. At that length a model stops being a forecast and becomes a way of arguing: every assumption has to survive being pulled on by someone with an opposing interest.

I frame the problem, build the multi-scenario model, stress-test the assumptions with Finance, Engineering and Legal rather than taking them on trust, and bring a recommendation to the VPs who have to sign it. When tariffs move or service turnaround slips, that model is what says how much it costs and what to do about it. I lean on AI tooling throughout, to compress the time between a question and a defensible answer.