I started by selling kites on a pavement.

The house was two hundred square feet and there were six of us in it. My father was in the army — a sepoy, not an officer, which in army housing is a distinction you feel every single day. The officers' children went to school in jeeps. We went in a battered truck that made me wonder what was holding it together.

For a while I resented him for that. It took me years to understand that he had given us the best nest he could, and that not everyone starts from the same place. That realisation is probably the single most useful thing I know.

He also taught me how to think, though neither of us would have called it that. He walked us to the bus stop every morning and told stories — mostly about the army, like the fortnight he spent unconscious after being wounded in the 1971 war. And he asked questions with no obvious answer. What would you do if you were walking and met an elephant? We would rack our brains the whole way to school. I still solve problems the way those walks taught me to.

I have always had to make my own money.

At fourteen I worked as an office boy in a bank. Coffee, stationery, dustbins, guiding people to the right counter. With the first real money I had earned I bought a calculator watch — which in the late eighties was roughly what an iPhone is now — along with new shoes and new clothes. I was left with a hundred and sixty-five rupees.

That hundred and sixty-five rupees is where everything else came from.

I bought kites wholesale and sold them retail. With those savings I bought Holi colours wholesale and sold them retail. With those savings, Diwali crackers. Whenever there was money sitting idle between seasons I lent it out in small amounts. Nobody had said the word microfinance to me; I had simply worked out that many small loans fail more gently than one large one.

At sixteen or seventeen, I bought twenty-five kilo sacks of tutti frutti, made them into small packets, and stapled ten of them to photographs of film stars. No wholesaler would switch suppliers for a scrawny boy, so I offered a deal nobody had offered them: pay me only for what you sell, return the rest. A wholesaler named Viresh said yes. That was the first time I understood that if someone will not give you business, you change the shape of the offer until refusing makes no sense.

Then I started running chit funds. The first was for a thousand rupees. Then two thousand, then five, then ten. By the end of Intermediate I was running two chits of twenty thousand rupees each.

I was not building a business. I was compounding, and I did not know that was the word for it either.

Helpdesk to Washington.

I joined Satyam as a contract helpdesk executive in 1998 and was permanent within nine months. Web designer, then developer, then team leader, then Singapore, London and the United States. In 2003 I went to the World Bank in Washington as an interaction designer, making software usable for people who found software hard to use, including people with disabilities. I met more nationalities under one roof than most people meet in a lifetime of travelling.

Funded, and then discouraged.

I came back and started NowPos in 2005, and alongside it TrulyIntelligent Technologies, the arm where the engine actually got built. We called it Lilo, after the character in The Fifth Element.

Lilo understood questions instead of matching keywords. You could ask how many airlines flew from the US to Kerala in a week, and it would understand what you were asking rather than hand you every page containing the words Kerala and airlines. We ran it on desktop machines. There were no GPU farms, no cloud, no AWS. This was 2005.

It worked. We raised a million dollars that year and landed on the front page of Business Standard. We had international patents filed and banks talking to us about putting it into ATMs. In 2006 Brand Capital, the Times of India Group's investment arm, came in with ten crore rupees in advertising equity at a hundred crore valuation. That same year we qualified at TREC and published a paper.

And then it stopped, for a reason that had nothing to do with the technology. Our investor was worried about Google. He did not encourage us to take it further.

I have thought about that for twenty years. The money was there. The technology was there. What was missing was the appetite to back an Indian company against a global giant — and that is still, I think, the thing that most holds us back. Not talent, not capital. Nerve.

What came after.

Possibillion, where we built the world's first automated idli machine and installed it in the Hyderabad Metro, until COVID closed it. Broadcast Wearables, where we made the world's first programmable LED t-shirt and ended up on History Channel. Now Neonflake, twenty engineers building AI and engineering products for clients. Some of it worked. Some of it did not.

Last month I filed five more patent applications, all of them in AI. That takes me to nineteen. I am still doing the thing I was doing in 2005, and this time nobody has to be talked into believing it matters.

And then I started painting again.

I never went to art school. As a child I sketched and painted constantly, and then responsibilities arrived and there was no time for it. The pandemic gave the time back. I have been painting since 2022.

I am aware this is an odd combination — patents and paintings. It has never felt odd from the inside. Both are the same activity: you start with nothing, you make something that was not there, and you find out afterwards whether it was any good.

What I still want

I want to walk on the moon.

There are two ways to get there. Become extremely fit, acquire the qualifications, and go to NASA. Or build an organisation that does the research and eventually makes the rocket. The second is more likely, and even if I never land, something gets left behind that someone else can build on.

I have hit rock bottom, and it is not a bad place to be. You can only go up from there.