I measure before I claim, and say so when a number is not defensible yet.

I started as one engineer on one repository. Today I lead applied-AI work at a consulting and product-engineering firm — building and directing across security operations, AI operations, sales tooling and physics-informed R&D — and I run VinaLabs, the studio where my own products live.

The through-line is measurement over assertion. I built Astram because I needed a system that could tell me what actually happened, not what I remembered happening. The same instinct runs through client work: correcting usage numbers against the production database before they reach an investor, or noticing that an AI eval suite measured whether the software behaved rather than whether the work was any good.

Vina (வினா) means question in Tamil. Every project here begins as one. The lab is where I try, measure and refine until there is an answer worth shipping — and where I say so plainly when there is not one yet.

What I am doing now

• Building. Astram Intelligence — the operating system I run my work on
• Leading. Applied-AI engineering and training at the firm where I work
• Teaching. A four-part course on disciplined engineering in the AI era
• Researching. Taste as a measurable thing — Paarvai for sight, Oli for sound

What I am like to work with

I lead with a recommendation, then the evidence for it. I name the gap before you have to find it. When a decision is yours to make, I will say so, give you my view, and not confuse the two. Status reports from me come as numbered lists, and they include what is not ready.

Read the ledger before you hire me. It is the most honest part of this site.