Platform Engineer / Tinkerer / Toolsmith
I am a platform engineer and a hobbyist tinkerer. I build dependable developer platforms and prefer simple tools over unnecessary complexity.
My curiosity about how the web worked pulled me into programming in 2008. I often wondered how a site like Google could serve the entire world, and why people typed google.com just to check if the Internet was up. Those were the days of Nokia PC Suite and ₹20 recharges to get the Internet working on my Zenith computer. I still remember being an early user of Google Chrome. I burned through three recharges just to download the browser and the Bluefish HTML editor (surprisingly, it’s still an active project), which I used to build my very first website.
At home, computer engineering books from my older sister were always around, and I read them out of curiosity. That exposure nudged me toward studying Computer Engineering. In college, I came across research on the Google File System and Borg. The scale and ambition of those systems fascinated me and drew me toward infrastructure: release automation, orchestration, and the often overlooked but essential parts that keep software reliable.
Since then, I’ve worked on large-scale migrations between cloud providers, built GitOps platforms that manage thousands of applications across many repositories, and spent time on the unglamorous work that keeps those platforms running: certificate rotation, staged production cuts, and the tooling around them. Cluster provisioning used to be a tangle: several teams each ran their own add-ons, in their own order, whenever a new cluster came up. I helped collapse that into one shared path so a cluster comes up with networking, platform add-ons, and GitOps wiring already decided. I also helped figure out Argo CD registration, so onboarding a cluster to the GitOps plane is a pull request instead of a manual handshake. I built an internal agentic platform where agents can work with services, sprint boards, and Slack the way a teammate would, and I have been putting that same idea into production: an on-call agent that reads Datadog, incidents, and GitOps state and posts a structured summary, and another that tracks Kubernetes release notes so software upgrade pull requests show up without someone chasing changelogs by hand. I enjoy turning messy operational challenges into clear, reproducible workflows that teams can rely on, so they can focus on building instead of fighting infrastructure.
I believe technology is most useful when it is widely accessible and scales gracefully. Personal hacks are fun; platforms that others can rely on are better.
Blog posts
- Building Production-Ready Monitoring Reports with Fast-Agent and Datadog MCP 2025/11/16
- How do I set up my Mac (dotfiles-first) 2024/08/24
- What I learned applying the AWS Well-Architected Framework 2023/08/22
- Kubernetes explained with car assembly 2019/04/17
- Auto-reload Development Mode — For celery worker using docker-compose and Django management commands. 2018/09/15