I'm Manan — a Senior Platform & Cloud Engineer who believes the best infrastructure work is invisible: things upgrade themselves, deploys don't page anyone, and new teams onboard without a meeting.
I've worked where cloud infrastructure meets scale. Not "we have three environments" scale — 150+ EKS clusters, ~1,000 nodes, 30K+ running pods, 500+ GPUs scale, where anything done by hand is already broken and every solution has to be a system.
My work splits into three layers: platforms teams build on (Terraform modules, GitOps addon delivery, self-service onboarding), fleet operations that keep those platforms current (cluster upgrades, bulk pipeline modernization), and observability that proves it all works (health sweeps, async service probing, post-change validation).
A fix applied to one cluster is a patch; applied to five hundred, it's a product. I build the second kind.
Blue/green by default, circuit-breaker rollbacks, idempotent tooling, ringed rollouts. Moving fast is a consequence of being safe.
Every fleet operation I ship produces an audit artifact. If it isn't in a report, it didn't happen.
The measure of a platform team is how rarely people need to talk to it.
Turning toil into products. Recurring infrastructure work becomes platforms — the kind of investment that shows up as faster teams and quieter pagers.
Fleet-scale operations. Upgrade orchestration, GitOps addon delivery, reusable deployment modules — built them, operated them, know where the bodies are buried.
Foundations that survive 10x. Multi-account AWS, CI/CD, and Kubernetes architectures designed so growth doesn't mean a rewrite.
Migrations and modernization. Monolith-to-microservices on ECS/EKS, console-clicks to Terraform, hand-deploys to blue/green pipelines — home turf.