My Story

Most weeks start the same way: a problem someone has lived with for years, and a question about which technology will fix it. My job is to slow that question down.

I head products, technology and AI for a group of schools, higher-education institutions and the businesses around them. I report to the promoter, sit in the board-level reviews, and lead a multidisciplinary team through five leaders: platforms, AI and finance systems, products, implementation and commercial operations.

The mandate is transformation. The method is older than any of the tools: understand the operating model first, then decide what to build, buy or leave alone.

Approval and audit checks across many institutions were slow, manual and hard to verify. Everyone assumed the software was the bottleneck.

It wasn't. The delay lay in the process and the evidence trail, not just the software. So we rebuilt both, and put AI where it could check the evidence rather than guess at it.

Days became minutes. With AI and evidence, not hype.

AI is most useful when nobody has to call it AI. Inside our products, models quietly tell urgent from routine, and a real lead from noise.

For a client, we built retrieval on their own knowledge, so the answers came from their documents, not from the internet's.

Building internal products taught me something: some of them were good enough to sell. So I built the case, the pilots, the revenue model and the P&L, and set up a software company inside the enterprise.

Today it has three product platforms, each live with pilot users. The first went from concept to paying customers through every unglamorous step: positioning, pricing, demos, proposals, onboarding.

It is early. That's the point: proof first, scale after.

I was a chemical engineer fresh out of college, standing in a control room where screens replaced walls and every number meant pressure, heat or money.

My work was automation and optimisation: making the plant's data tell the truth, then using it to run the units better. It was my first lesson in diagnosis. The problem is rarely where the alarm is.

Next came a global R&D lab and a different kind of chemistry: what happens to fabric, fragrance and water in homes across the world.

I turned global laundry science into products made for Indian factories and Indian homes, and helped build a global R&D capability in Mumbai, capturing knowledge that had lived in another country and making it work here.

At a global beverage company I was picked for a small global team with an unusual brief: invent a new kind of beverage dispenser, from zero.

It was a four-year programme and every stage of it was ours: sitting with retailers in India and China to understand what they needed, tearing down what already existed, then writing the requirements, the technology roadmap, the design, the engineering, validation, manufacturing and pilots.

The answer was a dispenser built to change with the seasons: hot in winter, cold in summer, modular in between. We connected it too, so it could report where it was, what it poured and when it needed help.

It was patented, and we took it to successful pilots. Inside it sat the part I am proudest of.

Beverage dispensers were expensive, largely because precise mixing depended on costly electronic valves.

But mixing is a fluid-dynamics problem before it is an electronics problem. I developed the fluid-dynamics basis for a Venturi mixing architecture using Bernoulli's principle, mathematical modelling and CFD. It replaced an approximately $500 component with an approximately $5 one.

Bernoulli did the rest.

What still stands: the dispenser testing specifications I wrote were adopted as global equipment-testing standards.

Then I moved from inventing to governing. I learned that innovation is a discipline, not a mood. Ideas move through gates, and each gate asks whether the idea still deserves the next rupee.

I governed the innovation portfolio for the Indian market and took global beverage innovation to the Indian shelf, with a gate at every step.

Then I left the corporate world to build something of my own: a blended-learning business, real teachers in real centres, joined to a digital platform.

It grew from a first centre to an app-first platform. Then a lockdown closed every centre.

The model's value was the teaching relationship, not the building. We moved operations online quickly, kept every member of the team and supported affected families.

The business was acquired, returning value to our investor.

After the acquisition I joined the acquirer to lead product: integrating the platforms we had brought together and building the foundation for a new product line.

Acquisitions are promises. Product is where you keep them.

Which brings me back to the Monday review. The questions are bigger now, and the tools are stranger: agents that act, models that need guardrails, tests and tools before they are useful at work.

The method has not changed since the control room. Understand the system. Find where it really breaks. Then build what lasts.