🏗️
Production Multi-Tier Serverless & Containerised Infrastructure
- Provisioned production-grade AWS infrastructure using Terraform across a single account with three distinct traffic flows
- Flow 1 — static content served globally via CloudFront backed by S3
- Flow 2 — dynamic API traffic routed through CloudFront to API Gateway, NLB, and ECS Fargate with RDS PostgreSQL as the transactional backend; Fargate Spot used for non-production to reduce compute costs
- Flow 3 — async event-driven processing via API Gateway publishing to SQS, with Lambda consumers writing results to DynamoDB, fully decoupling async workloads from the synchronous API layer
CloudFrontS3
API GatewayNLB
ECS FargateRDS PostgreSQL
SQSLambda
DynamoDBTerraform
🌐
Hybrid Multi-Account Network Architecture
- Architected secure hybrid connectivity bridging on-premises infrastructure with AWS using a Site-to-Site VPN terminating into a shared services account
- AWS Transit Gateway acts as the central routing hub, distributing all on-premises traffic across UAT and Production VPC attachments via TGW routing tables
- Single VPN connection serves all environments, eliminating the need for separate tunnels per environment
- Hub-and-spoke model enforced clean environment isolation while providing a single point of network governance for all cross-environment traffic
Site-to-Site VPNTransit Gateway
Shared ServicesUAT + Prod VPC
Hub & SpokeTerraform
💰
AWS Cost Optimization & FinOps Automation
- Configured S3 lifecycle policies to automatically transition and expire objects across storage tiers
- Migrated all non-production EKS workloads to Spot Instances, reducing compute costs while maintaining availability
- Built a Python cleanup script triggered via EventBridge to automatically remove orphaned EBS volumes and stale snapshots on a schedule
- Replaced 3–4 manually managed EC2 instances behind an ALB with an Auto Scaling Group, improving resilience and eliminating manual overhead
- Used AWS Compute Optimizer to identify and implement rightsizing recommendations across EC2 instances
- Exported Cost Explorer recommendations, mapped against existing workloads, modelled estimated savings, presented to client, and facilitated purchase of Savings Plans and Reserved Instances
S3 LifecycleSpot Instances
EventBridgePython
ASGCompute Optimizer
Savings PlansReserved Instances
⚡
Serverless Attendance Tracking Application
- Built a fully serverless attendance tracking application, decoupling frontend from backend for high scalability with zero server maintenance
- Provisioned frontend hosting using S3 and CloudFront for fast, globally distributed, and secure delivery
- Backend powered by API Gateway, Lambda, and DynamoDB to handle all dynamic requests without any server provisioning
- Automated infrastructure deployments via a Dockerized Jenkins CI/CD pipeline with Git hooks triggering Terraform plan on every commit and manual approval gate before apply
S3CloudFront
API GatewayLambda
DynamoDBTerraform
Jenkins CI/CDDocker