backend · cloud platforms · golang
I like finding out why things broke, fixing them properly, and building systems that scale and stay up. Outside of that: new tech to poke at, coffee, and football.
I'm a backend engineer at IBM, working on the cloud layer that database products run on. Day to day that means writing and maintaining Kubernetes controllers, helping database teams get their products running on the platform, and taking my turn on the pager. It's the boring-on-purpose infrastructure that other teams ship on. That's exactly what I find interesting about it.
Before this I worked on production Go microservices: EDI pipelines, a private blockchain on Cosmos SDK, and on-demand AWS environments. Earlier I helped keep the backend of India's largest voice-data collection platform running while looking after GPU infrastructure for its ML teams. Small teams taught me to ship fast; platform work is teaching me the discipline to ship things that last.
Distributed systems are the part I keep coming back to. Something about how failure emerges from machines that all individually did the right thing never stops being interesting. Outside of work I root boxes on HackTheBox. Understanding how things break is useful regardless of which side of the problem you're on. When I'm away from the screen it's football or coffee. Always happy to chat.
v3.x 2026 — present ● in production
v2.x 2023 — 2026
v1.x 2021 — 2023
v0.x pre-2021
// the short list of things I use enough to be opinionated about
module github.com/ashishkhuraishy/engineering
go 1.24
require (
golang vDaily // services, controllers, CLIs
kubernetes vDaily // custom controllers, operators, debugging at 2am
grpc-and-rest vProd // api design, protobuf, versioning
postgres / redis vProd // state, caching, "why is this slow"
rabbitmq / queues vProd // async pipelines, exactly-once-ish
docker / ci-cd vProd // images, pipelines, releases
observability vProd // grafana, prometheus, loki, slo thinking
aws / cloud-infra vProd // provisioning apis, automation
ml-infra vShipped // gpu clusters, validation pipelines
new-shiny-things vAlways // currently learning: see /projects
)
Versions are honest: vProd means I've run it in production and
been paged for it.
Go CLI that monitors RAM and CPU across multiple servers and renders the numbers as live terminal graphs. Built so I could watch my whole fleet without keeping five SSH tabs open.
Docker base image for the Kaldi speech toolkit, optimised for running ASR servers on top of it. 2k+ pulls on Docker Hub from people who also didn't want to compile Kaldi from scratch.
The page you're reading, built with Astro, Tailwind, and an animated request trace. Permanently marked "building" because that's the honest status of every personal site ever made.
Happy to talk platform problems, Go, infrastructure, AI systems, or why your cluster is doing that.