The problem I work on
Big companies have architecture review boards. A non-technical founder has an agency salesperson. Every week, founders commit tens of lakhs — sometimes their whole runway — on technical decisions they have no independent way to check: what to build first, who should build it, whether the quote in front of them is honest, whether the idea needs AI at all. The information exists; what's missing is judgment that sits on the founder's side of the table. That's the gap this site exists to close.
Where the judgment comes from
I've been building software for 18 years, across four full technology waves — network security at Juniper and Brocade, software-defined networking, cloud-native SaaS, and now production Gen AI. Every wave I followed for the same reason: that's where the undiscovered territory was.
In the middle of that arc I switched chairs. I founded Naanal Technologies, a logistics SaaS, and ran it for 5 years — 0 to 45 people, 0 to ₹6 crore ARR. Those years are why this site sounds the way it does. I've made payroll decisions and scope-cut decisions in the same week. I've hired the first engineer, read the agency quotes, and shipped the version one that was embarrassing but sold. The decisions in these guides aren't theory to me; I've paid for the wrong answers personally.
What I do now
Today I'm a Principal Architect running Gen-AI SaaS at scale — a 14-person engineering team, LLM systems at 50 million tokens per minute in production. Alongside that, I'm advising early-stage founders on product and technical decisions: the recurring technical decisions of going 0 to 1, from “is this worth prototyping” to “who builds it” to “when do I hire someone who replaces me in this role.”
LinkedIn named me a Top AI Voice in 2024, which mostly means one thing to you: I write about these decisions in public, in plain language, every week.
Why 0 to 1, and why exit by design
I like the earliest stage because it's the stage where judgment is worth the most and process is worth the least. One good decision at 0-to-1 — the right scope, the right builder, the honest read on whether AI belongs in the product — moves more than a quarter of optimization later.
And I design every engagement to end. I take a founder from idea to shipped product to their first engineering team, then hand off to their full-time technical leader and leave. Dependence is a bad product. Success, for me, is you no longer needing me — the long version of that is on the Method page.
Find me
The fastest way to reach me is a LinkedIn message. The newsletter is where new decision guides land first.
My engineering research archive lives in the Lab.