About
I keep large, latency-sensitive systems alive — and increasingly the GPU and compute platforms that AI runs on. Seven years in infrastructure and reliability engineering, from bare-metal datacenters to production Kubernetes at scale.
At Booking.com I define SLIs, SLOs, error budgets and alerting for hundreds of services across more than 20 engineering teams. I run production Kubernetes (EKS) with autoscaling and topology-aware scheduling, write the Terraform and CDK that new environments are built from, and chair the review board that follows up after every outage.
Before that I ran more than 150 Kubernetes clusters across bare-metal hardware and cloud, carrying real-time video inside roughly a 40ms latency budget. At that number nothing works unless it is automated. That habit — finding where the milliseconds and the capacity actually go — is what I am bringing to AI infrastructure.
Experience
— Present
I own SLIs, SLOs, error budgets and alert design for hundreds of high-throughput services across more than 20 engineering teams. I run production Kubernetes (EKS) with autoscaling, multi-AZ resilience and topology-aware scheduling, and author the Terraform modules and AWS CDK stacks that new environments are built from. I lead incident response on an on-call rotation and chair the corrective-action review board that turns outages into permanent fixes. I also cut deployment time by 30% by restructuring CI/CD into concurrent and downstream stages, and built AI-generated service documentation that saves around 10 engineering hours per team.
- Kubernetes (EKS)
- Terraform
- AWS
- SLOs & Observability
- Incident Response
- ArgoCD
— July 2022
As a one-person team I migrated the entire company stack from a monolithic, datacenter-hosted setup to containerised Kubernetes (EKS, AKS) across AWS and Azure, databases included. I authored the AWS CDK stacks and Terraform modules that became the organisation's IaC standard, built CI/CD delivering through ArgoCD into dev, staging and production, and trained the developers on containers, Kubernetes and Helm.
- AWS
- Azure
- Kubernetes
- AWS CDK
- Terraform
- Migration
— Aug 2021
I provisioned and operated more than 150 Kubernetes clusters spanning on-premise bare-metal datacenter hardware and cloud (EKS, AKS), carrying real-time video inside roughly a 40ms latency budget. At that scale everything has to be automated: provisioning, lifecycle, and node and agent health checks feeding remediation. I managed the whole cloud, network, security and compute estate as Terraform code and authored the reusable module library, built the monitoring and logging stacks on Datadog, New Relic, CloudWatch and Sumo Logic with dashboards-as-code, delivered an AWS cost-optimisation programme, and set up Okta SSO with least-privilege IAM.
- 150+ K8s Clusters
- Bare Metal
- Terraform
- Datadog
- Low Latency
- Cost Optimisation
Projects
Portfolio
Portfolio is my personal portfolio, where you can see my projects and contact me.
- Next.js
- TailwindCSS
- JavaScript
