the story

I build things, and I write to understand why people do what they do. For a long time I thought those were two separate habits. They are not. Both come from the same place: a childhood where reality was heavy up close, and thinking at scale was the only way to make it feel survivable.

I'm Malav Shah, based in Bangalore. Most days I'm building an AI-native creative production studio. The bet is simple: AI has collapsed the cost of making creative work, so the scarce thing is no longer production, it's judgment. It's what an agency looks like when machines do the volume and humans own the taste, so consumer brands get campaigns that used to take a team of ten and a month, in a fraction of the time and cost. I build it with my collaborator Priyanshi and a small team.

That studio is the front of a wider set of bets, and they all rhyme.

One is a consumer tool that lets people with the most interesting problems to solve, small business owners and operators in markets that software forgot, talk their way to a real product in their own language, instead of learning to code or hiring someone who can. It's a bet that the next billion builders won't type.

Another is infrastructure for the opposite end of the spectrum. As AI agents start touching money, someone has to answer for what they did and whether it was allowed. This is the audit trail, the guardrails, the regulatory scaffolding that lets a serious institution actually deploy an agent.

And there's a children's IP, which exists partly because the stories we absorb as kids are the first belief-forming systems we ever meet, and partly because it's a joy to make something that isn't optimized for a funnel.

If there's a single thread under all of it, it's this: I'm obsessed with how beliefs form and how systems shape behavior. Ad creative is belief formation at six seconds. A market thesis is belief formation at the scale of capital. A children's story is belief formation before you have any defenses against it. A compliance layer is a system deciding what an agent is allowed to believe it can do. I keep building in different rooms of the same house. the six seconds is not a figure of speech. it is roughly how long you have before the thumb moves.

Before all this, I worked on consumer AI products (Immersfy, Blix), did a stretch in fintech consulting, and spent a lot of time making content, which taught me more about human attention than any framework did.

The writing is the other half. I keep a Substack called The Understory: quiet essays on identity, intimacy, internet culture, and the emotional systems running underneath ordinary life. It is deliberately not about my companies. It's where I think without a deliverable at the end, and it's honestly where a lot of the sharper product ideas start before I know they're product ideas.

I read and think in the register of people like Nikhil Kamath and Chamath Palihapitiya, not for the theatre, but for the discipline of naming a structural shift, finding the non-obvious wedge, and staking a real position. Much of what I publish under Theses is me trying to earn a point of view rather than borrow one.

So: builder on the loud days, essayist on the quiet ones, and underneath both, someone trying to understand the invisible machinery that decides what people believe and do. This site is where those two people finally share an address.