Kazi Shiplu
Lead Engineer. I build AI agents that survive production - and I can show you what breaks. 8 years of backend and platform work underneath.
01
Track record
end to end, in productioncat ~/experience
| role | Lead Engineer · backend, platform & AI | current |
| operations | I deploy, monitor, and get paged for my own systems | on-call |
| open source | blip · chronostash | maintainer |
| homelab | 3× Proxmox · dual-WAN · local AI | self-hosted |
| years shipping | 8 |
02
Stack
the tools I ship with- Node.js
- TypeScript
- NestJS
- PostgreSQL
- Redis
- RabbitMQ
- Docker
- Kubernetes
- Terraform
- Cloudflare Workers
- Proxmox
- Local LLMs
03
Blip
uptime monitoringAn uptime monitoring platform that runs entirely on one Cloudflare Worker with D1. No servers, no containers. It watches my clients, my projects, and my homelab.
04
ChronoStash
backups that restoreScheduled, encrypted backups for Postgres, MySQL and MongoDB to any S3-compatible storage, with restore that actually works. Built after learning the hard way what an untested backup is worth.
05
The homelab
where everything gets proven first3× Proxmox nodes, MikroTik dual-WAN with real failover, and a local AI stack running open models on my own hardware, exposed to the internet through Cloudflare Tunnel without opening a single port.
- Proxmox
- MikroTik
- Cloudflare Tunnel
- WireGuard
- local LLMs
06
Writing
what actually breaks07
Straight answers
the questions people ask first- Who is Kazi Shiplu?
- I'm a Lead Engineer based in Dhaka, Bangladesh, with 8 years of backend and platform work behind me. Today I design and run AI agent systems in production, and the infrastructure they depend on. I operate what I build, which means I'm the one on call when it breaks.
- What is Kazi Shiplu working on now?
- Agent systems that have to hold up against real traffic: permission-aware retrieval, idempotent tool calls, and the failure handling underneath both. Alongside that I maintain two open-source tools, Blip for uptime monitoring and ChronoStash for database backups, and a three-node homelab where every architecture I write about runs before I recommend it.
- Does Kazi Shiplu take consulting or contract work?
- Yes, for work on production systems: putting an agent in front of real data, or fixing the backend and operations underneath one that is already live. A LinkedIn DM is the fastest way to start, email is second. Both are at the bottom of every page.