THE
WORK

Six weeks. Six projects. From CI/CD pipelines to Kubernetes observability to serverless event pipelines. Every project is fully deployed, documented, and tested. Every blog post walks through the architecture, the debugging, and the lessons learned.

06
Projects
06
Blog Posts
03
AWS Certs
100%
Shipped
01

AWS CI/CD
PIPELINE WITH
BLUE/GREEN

TYPE: AutomationSTACK: AWS + TerraformLOC: ~800

Full CI/CD pipeline on AWS with zero-downtime blue/green deployments. CodeCommit triggers CodeBuild, which creates an artifact for CodeDeploy. Manual approval via SNS before production rollout. Tested rollback by breaking the health endpoint.

“The moment the blue/green rollback worked exactly as designed — that's when DevOps clicked.”
/ TECH STACK
CodePipelineCodeBuildCodeDeploySNSTerraformNode.jsBun
📝 BLOG POST →⚡ GITHUB →
02

THREE-TIER
AWS
ARCHITECTURE

TYPE: InfrastructureSTACK: ECS + RDS + RedisLOC: ~1500

Production-grade 3-tier architecture — ECS Fargate, RDS PostgreSQL, ElastiCache Redis in a custom VPC. Full CRUD API with cache-aside pattern. 6 Terraform modules, 37 resources.

“Built the VPC from scratch because I wanted to understand every subnet, every route, every NAT gateway.”
/ TECH STACK
ECS FargateRDS PostgreSQLElastiCacheALBVPCECRTerraform
📝 BLOG POST →⚡ GITHUB →
03

GITOPS ON
EKS WITH
ARGOCD

TYPE: KubernetesSTACK: EKS + Helm + ArgoCDLOC: ~2500

Full GitOps workflow on EKS. Push to main → ArgoCD auto-syncs. Custom Helm chart with 8 templates. PostgreSQL StatefulSet with PVC, Redis, AWS Load Balancer Controller, EBS CSI Driver, IRSA for pod-level IAM.

“ArgoCD showed Unknown sync status for 20 minutes because of a single misindented line in a Helm template.”
/ TECH STACK
EKSHelmArgoCDIRSAALB ControllerEBS CSITerraform
📝 BLOG POST →⚡ GITHUB →
04

PROMETHEUS +
GRAFANA +
ALERTING

TYPE: ObservabilitySTACK: Prometheus + GrafanaLOC: ~1200

Full observability stack on EKS. Instrumented Node.js API with 7 custom Prometheus metrics. Built a 9-panel Grafana dashboard and wrote 9 custom alert rules. Deployed via GitOps using kube-prometheus-stack.

“Going from 'we use Grafana' to writing PromQL queries and designing alert thresholds from scratch.”
/ TECH STACK
PrometheusGrafanaAlertmanagerPromQLServiceMonitorprom-client
📝 BLOG POST →⚡ GITHUB →
05

LOKI +
PROMTAIL
LOG AGG

TYPE: ObservabilitySTACK: Loki + PromtailLOC: ~600

Added log aggregation to the observability stack. Promtail DaemonSet ships logs to Loki, queryable via LogQL. Built a metric-to-log correlation dashboard — click a spike, see the logs from that exact time range.

“Metrics tell you that something is wrong. Logs tell you why. You need both.”
/ TECH STACK
LokiPromtailLogQLGrafanaDaemonSetArgoCD
📝 BLOG POST →⚡ GITHUB →
06

SERVERLESS
EVENT
PIPELINE

TYPE: ServerlessSTACK: Lambda + DynamoDB + SQSCOST: $3.60/mo

Real-time event-driven data pipeline on AWS. 4 Python Lambdas, API Gateway HTTP API v2, DynamoDB with GSI, SQS with DLQ. Fire-and-forget async processing, atomic counters, conditional writes, partial batch failure. $3.60/month vs $213 for EKS.

“Same architectural complexity. 60× cheaper. Serverless wins for event-driven workloads.”
/ TECH STACK
LambdaAPI Gateway v2DynamoDBSQS + DLQPython 3.12Terraform
📝 BLOG POST →⚡ GITHUB →🎮 LIVE DEMO →