AWS platform engineering work — real projects, real numbers, no fluff.

📊 By the numbers

Across the projects below:

• 💸 $90K+/year saved in AWS costs through targeted optimisation
• ⚡ 4,000+ messages/second handled at 99.9% SLA across isolated AWS accounts
• 🚀 16x faster deployments after replacing CDK/CloudFormation with Terraform
• 💸 60%+ Kubernetes cluster cost reduction via Karpenter + Spot instances

🗂️ What’s covered

If any of these sound familiar, there’s a project below that’s relevant:

• 📦 On-prem to AWS migrations — phased, zero downtime, full IaC, product keeps shipping
• 🧩 Event-driven platforms — MSK (Kafka), cross-account IAM, PII-compliant, audit-ready
• 🌍 Multi-account AWS platforms — EKS, ECS, multi-region, end-to-end CI/CD
• 🔒 Security & compliance — CloudTrail, GuardDuty, Security Hub, IAM least-privilege
• 💸 Cost optimisation — from right-sizing to tracing a $30K/year overspend to a single misconfiguration

🏗️ Projects

📦 On-prem to AWS migration

🚀 16x faster deployments 💸 $60K/year saved ⏱️ Zero downtime

An aging on-prem environment moved to AWS in phases — no downtime, no disruption. Replaced CDK with Terraform and cut deployment time 16x. Automated scale-to-zero on non-prod environments saved $60K/year.

🧩 Kafka-based Enterprise Service Bus

⚡ 4,000+ msg/s 💸 $30K/year saved in 5 days 🔒 99.9% SLA

AWS MSK event bus serving 20 teams and 300+ supermarkets across fully isolated accounts. Maintained 99.9% SLA with automated alerting and self-service access via Jira. Traced a GuardDuty overspend to a single misconfiguration and cut costs by $30K/year in 5 days.

🌍 Social Media Platform

💸 60%+ cluster cost reduction 🌍 4 environments, 15+ microservices

Multi-account AWS platform for a global social app — 4 environments, 15+ microservices, multiple regions, all reproducible via Terraform. Karpenter + Spot instances brought cluster costs down by 60%+.

🤔 Got a similar problem? Send me a message — I’ll tell you within 24 hours if I can help.