Layak Singh Thinker and doer. I work on hard ground, and stay until it works.

The businesses I choose all have the same shape. A parent in a district town pays a tutor in cash and will not pay twice for something that did not work. An insurer has a stack of claims on a desk and no way, from data, of telling which one to look at twice. A pandit performs a puja and ends the week with nothing that counts as a record of the work. None of that is a software problem. These markets run on trust, cash, people and physical ground, and they get left alone because the slow parts are verification and distribution, and because somebody at the far edge of it has to be paid on time. I have been starting companies since IIT Kharagpur, where the first was research writing sold between classes to anyone who needed it written. That one proved money turns up if the work is good, and I have spent every year since finding out how much else has to be true before it does.

Three of mine closed. dateIITians was a network for IIT and IIM alumni that I could not grow. Cogxio, the hyperlocal product it turned into, shut in 2016. LiveOnGo ran beside it and went the same way. The matching worked in all three, which was exactly the problem. I had built the part I was good at and had no answer for how the next ten thousand people were meant to hear about it, or why they would pay once they had. Four years to learn that, and it is still the most useful thing I know. What you can build is almost never the constraint. Reaching the buyer is, and after that it is whether the buyer will trust you with money.

Artivatic.ai came out of that in 2018, on a bet that was commercial before it was technical: an Indian insurer would let a machine make a decision it had always made by hand, provided the decision arrived as one step it could change rather than a system it had to replace. We built underwriting, fraud and claims for companies that had none of it. What got a signature was one property of what we sold: a buyer could put that step back the way it had run on Friday and lose nothing, and nobody inside a regulated company makes a larger decision quickly. RenewBuy acquired the business in February 2022 and I stayed. From 2023 I took the platform outside India, into South Africa, the United States, the Middle East and the Philippines, on the plain reading that the paperwork it replaces looks much the same in each of those markets. On 31 August 2026 RenewBuy and InsuranceDekho came together under Artivatic as the holding company, the group now trades as InsuranceDekho, and I left the executive role that day, eight years after founding it, staying on as an advisor to the group. Four and a half of those years came after the sale. Most founders are gone long before that.

DharmikVibes is what I am building now, and it is the same job in a sector nobody has run as a business. Several hundred million people organise their year around temples, and nearly all of the money in it moves as cash and word of mouth, so there is no record anywhere a bank would look. The unglamorous part is the whole thing. A pandit has to be verified to a standard that would satisfy a family elder. A temple's own rules and dates have to be right before anybody books against them. Then the smallest partner has to be paid on the day he was told, because that is what decides whether he is still doing this next year. The temples and the people who serve them are in India and stay there. The devotees are wherever they have settled, so someone living overseas can sponsor a puja in their own name at a temple they cannot reach and have prasad sent on. Alongside the company I advise RODIC Consultants in infrastructure, advise the Artivatic group, and take on a small number of early-stage founders and non-profits. I write as well, usually while I am still working out what I think.