SHIVR.DEV

$ cat tagline.txt
blog
scroll down▼
$ dashboard --metrics
┌─git-stats─┐
[─]
$ git log --oneline | wc -l
> syncing from github...█
┌─ai-usage─┐
[─]
$ ai-metrics --lookback 30d --summary
> syncing from providers...█
┌─languages─┐
[─]
$ tokei --sort lines
> fetching language data...█
$ ls -la ~/projects
┌─projects─┐
[─]
STNAMELAST
GoTypeScriptReactConvexPython
$ cat ~/about.md
┌─about.md─┐
[─]
# Shiv Rossi
> I'm Shiv Rossi, a Staff GenAI engineer focused on agentic systems, evals, and reliability.
> Agents are easy to make look impressive. It is much harder to make them reliable, useful, and worth deploying once users, workflows, and incentives get involved. I care about what breaks in real AI products, what matters operationally, and how to build systems that survive contact with reality.
> My background is in ML infrastructure and data systems — production platforms at Meta, early engineering at AI-native startups, and a lot of time spent on what happens after the demo stops working. I write about what I learn the hard way.
$ cat ~/.contact
┌─contact─┐
[─]
$ echo $EMAIL
┌────────────────────────────────────────┐
│  built with minimal human involvement  │
└────────────────────────────────────────┘
shiv rossi © 2026 — all rights reserved