The Question That Had No Next Step

For years, I measured my direction by one exam. I sat it six or seven times, while working at a CA firm, in procurement and operations, sometimes with no job at all. Every year, the same hope. It never worked out.

Eventually I stopped forcing a path that clearly wasn't working and moved to Bengaluru, into a life where nothing was familiar. Somewhere in that stretch, I found technology.

Nobody assigned it to me. No exam, no deadline, no one checking if I understood it. I was simply curious. I wanted to know how things worked, and I wanted to build them myself.

Then I hit the obvious wall: I didn't know enough. For a while, that became maybe this isn't for someone like me. But something kept pulling me back. Every time I sat down to build, I wanted to come back the next day. That was the signal I trusted, before I trusted anything else.

I stopped asking do I know enough and started asking what can I figure out right now.

The first question has no next step. You just wait to feel ready. The second one has an action attached immediately.

That's the part I acted on. I'd taken a role hoping it would turn into something more, a way in, eventually, to the work I actually wanted. It didn't open that way. Most days felt like something I was pushing through, not showing up for. People around me had opinions. It's not easy. It's a stupid decision. You can't just do this.

I didn't have a counter-argument, not really, just a quiet certainty I couldn't yet prove, and needed to go test.

I left my job first, with nothing lined up. Then I went back to basics, actual basics, before I let myself think about building anything real. Third Wave became my office. Same table, most days, learning syntax and concepts I should've probably known already. Only once that foundation held did I start taking on things that needed more, new topics, harder problems, work that stretched past what I already knew.

Learn. Build. Get stuck. Figure it out. Build again.

I once spent a hackathon building an AI agent under a hard deadline, sure by hour six that my whole approach was broken, until I found the one piece of context I'd forgotten to give it. It placed in the top 400 globally, out of thousands. Later, I built a multi-agent system that cut incident diagnosis time by 95%. Later still, alone, an AI tool with no team and no roadmap but my own. I rarely knew the answer. I knew how to keep looking.

That's still how I work. I'm building a startup now, still open to what comes next, still the least experienced person in most rooms I walk into. I recently submitted an application to Y Combinator's Fall 2026 batch, with no idea if it goes anywhere. I did it anyway, the same way I did most things that mattered, before I felt ready for them.

I don't count those six or seven attempts as time I lost. They're where I first learned to tell the difference between something that needed more effort, and something that was never mine to begin with.

I don't know how to do everything I'll need to know. But not knowing isn't where the story ends.