Vivek Singh

Building Public AI Product for Tech

About

I design and ship AI products for technology teams. My work sits where product, LLM architecture, and production systems meet: application development, cloud, and DevOps. I care about tools that stay model-agnostic, run locally or in a real environment, and survive the jump from prototype to production. Right now I am building multiple public AI products for the community - starting with Cosen (cost, observability, and security for any LLM) - and more in the same line.

Badges

Tastemaker

Gone streaking 10

Gone streaking

Gone streaking 5

Maker History

Deadpathfind the code your agents keep rewriting around
Sep 2026
CosenLLM-agnostic cost, traces, and security. Self-host it.
Sep 2026

🎉

Joined Product HuntJanuary 4th, 2026

Forums

•

28d ago

Deadpath - find the code your agents keep rewriting around

Deadpath finds dead code in your repo unused exports, orphan files, unreachable functions, and dependencies that nothing imports—then puts that scan where your coding agents work: Claude Code, Cursor, and Codex. Most “AI dead code” tools ask a model to guess what looks unused. Agents then rewrite around the same dead files week after week. Deadpath flips that. The core scanner is deterministic. It builds a reference graph from the codebase and reports high-confidence findings with evidence.

•

2mo ago

Cosen - LLM-agnostic cost, traces, and security. Self-host it.

Cosen — local-first, LLM-agnostic cost, observability, and security for AI products.