About

I'm Waleed, a platform engineer and solutions architect based in Islamabad, Pakistan 🇵🇰.

I came to infrastructure the long way, through the product. I started full-stack, building desktop software, web apps, backend services, frontends, and the cloud infrastructure underneath them. Over time, that underneath part kept pulling me in: distributed systems at scale, deployments, reliability, observability, CI/CD, Kubernetes, all the machinery around an app that decides whether it survives contact with production.

That became the main thread.

At Aurora and Stakater I worked across the stack but kept moving deeper into platform engineering, migrating applications onto Kubernetes, building OpenShift-based platforms, writing operators and automation, and helping teams ship without hand-rolling the same infrastructure every time.

These days it's mostly Kubernetes at scale: multi-cluster networking, node lifecycle, GitOps, platform automation, and the unglamorous production details that separate a platform demo from a platform people can actually run.

I'm currently at Kubermatic, building open-source infrastructure tooling and designing Kubernetes environments for enterprise customers across cloud, on-prem, air-gapped, and multi-cloud setups. On the product side, my main focus is KubeLB, a centralized load-balancing platform for multi-cluster Kubernetes.

Like everyone with an internet connection, I've been pulled into the AI gravity well too, though I'm mostly interested in the unglamorous layer underneath it: GPUs, inference, model serving, training workloads, routing, scheduling, and the real cost of running models yourself instead of renting them forever. So I built Kubernetes LLM Inference Platform, a GitOps-managed stack for self-hosting inference on your own GPUs.

I use this stuff daily too, not just build it: agents, AI-assisted everything, and an AI bill that's quietly overtaken my food delivery.

This site is where I write about what I'm building, learning, and occasionally overthinking.