About

I hold the work to a standard. Then I publish the standard.

I write and teach because production systems punish sloppy thinking, and most AI tools make confident mistakes easier. This site is that public practice — 17 years of it — not a job title.

The canonical artifact is github.com/arunveersingh/ai — 11 agent skills that will say no, not ready, and do not proceed. That is the same bar I use on architecture: unverified assumptions do not ship.

I teach in English on Arun | Engineering Judgment and in Hindi on Developer Notes Hindi. Selected talks on this site are about judgment — first principles, agent pipelines, inference, seniority. Syntax drills stay in the library. Written courses are listed here.

Writing lives here first, then LinkedIn. Selected systems work: builds — MCP, SBOMs, protocol and architecture labs — not every repo I ever used to learn.

Reach me at arunveersingh9@gmail.com. New essays go out over RSS.

Employment history is LinkedIn. How I think is this site.