ls projects/

Production-grade distributed systems, government-grade security, and cloud-native solutions.

Infrastructure as CodeGitOpsObservabilityOPEN SOURCEOPEN SOURCE

AWS Cloud Platform — EKS Infrastructure with Terraform & GitOps

Production-grade cloud infrastructure: IaC, GitOps, and full observability on AWS

Production-grade cloud infrastructure on AWS, fully provisioned with Terraform. Features EKS Kubernetes cluster, RDS PostgreSQL, ElastiCache Redis, and Application Load Balancer across multiple availability zones. Automated deployment pipeline with GitHub Actions pushing to ECR and ArgoCD syncing to EKS. Complete observability stack with Prometheus, Grafana dashboards (versioned as code), and Loki for centralized logging. Multi-environment setup (dev/staging/prod) with Terraform workspaces.

$ terraform apply --target=aws_infra

Terraform
IaC Provisioning
GitHub Actions
CI Pipeline
Amazon ECR
Container Registry
ArgoCD
GitOps CD
AWS VPC — Multi-AZ
Route 53
DNS
ALB
Load Balancer
EKS Cluster
svc-1svc-2svc-3svc-n
RDS PostgreSQL
Multi-AZ
ElastiCache Redis
Cluster Mode
S3
Terraform State
Prometheus
Metrics
Grafana
Dashboards as Code
+
Loki
Log Aggregation
dev / staging / prod — Terraform workspacesBlue/green deployments via ArgoCD

Infrastructure Metrics

Infrastructure
100% IaC
Deployment
<5 min
Environments
3
Monitoring
Real-time

Infrastructure as Code

  • Terraform modules — VPC, EKS, RDS, ElastiCache, ALB
  • Multi-AZ deployment — high availability by default
  • Remote state — S3 backend with DynamoDB locking
  • Terraform workspaces — dev / staging / prod
  • IAM roles & policies — least-privilege access
  • Security groups — network isolation per service

GitOps Pipeline

  • GitHub Actions — build, test, push to ECR
  • ArgoCD — declarative CD, auto-sync to EKS
  • Blue/green deployments — zero-downtime releases
  • Slack notifications — pipeline status alerts
  • Image scanning — Trivy security checks in CI
  • Helm charts — versioned Kubernetes manifests

Observability Stack

  • Prometheus — metrics collection & alerting rules
  • Grafana dashboards — versioned as code in Git
  • Loki — centralized log aggregation from all pods
  • Alertmanager — PagerDuty & Slack routing
  • Kubernetes events — cluster health monitoring
  • SLO tracking — availability & latency targets
AWSTerraformKubernetesEKSArgoCDGitHub ActionsPrometheusGrafanaLokiDockerHelmECR
Monorepo to Polyrepo5 Independent ServicesNetflix PatternOPEN SOURCE

Cloud-Native E-Commerce Microservices Platform

From Monorepo to Polyrepo: Production-Grade Distributed Systems Evolution

Enterprise-grade microservices architecture demonstrating real-world architectural evolution from monorepo to polyrepo patterns. Built with Spring Boot and Spring Cloud, this project showcases the journey from a single-repository architecture to independently deployable services—the same architectural pattern used by Netflix, Amazon, and Uber. Features 5 independently deployable microservices with service discovery, centralized configuration, API gateway with circuit breakers, OAuth2 authentication, and individual CI/CD pipelines per service.

System Architecture (Excalidraw Style)

ClientHTTPSCloud GatewayPort: 9090OAuth2 + Circuit BreakerAuth0OAuth2 ProviderConfig ServerPort: 9296Product ServicePort: 8081RESTOrder ServicePort: 8082RESTPayment ServicePort: 8083Eureka ServerPort: 8761Service DiscoveryproductDbMySQLorderDbMySQLpaymentDbMySQLLegendHTTP RequestConfig/DiscoveryService-to-ServiceDatabase ConnectionAPI Gateway / AuthInfrastructure ServicesBusiness ServicesDatabasesExcalidraw style

Architectural Evolution

V2: Polyrepo Architecture

Independent repositories per service following Netflix/Amazon patterns. Each service can be deployed, scaled, and maintained independently.

Polyrepo5 ReposIndependent CI/CDK8s ReadyNetflix Pattern
Benefits
  • + Independent deployments per service
  • + Team autonomy and ownership
  • + Isolated failure domains
  • + Technology freedom per service
  • + Parallel CI/CD pipelines
  • + Scalable per service
Trade-offs
  • - More complex initial setup
  • - Requires service discovery
  • - Need contract testing
  • - Distributed configuration
Performance Comparison
Build Time
15 min3 min
Deploy Time
20 min5 min
Blast Radius
100%20%
Rollback
FullPer Service

Microservices Architecture

  • Service Discovery - Netflix Eureka
  • API Gateway - Spring Cloud Gateway
  • Circuit Breaker - Resilience4j
  • Config Server - Git-backed storage
  • Inter-service - OpenFeign
  • Database per Service - MySQL

Architectural Patterns

  • Polyrepo - Independent repos per service
  • API Gateway Pattern
  • Saga Pattern - Distributed transactions
  • Config Externalization
  • Service Mesh Ready
  • OAuth2 + JWT Authentication

DevOps & CI/CD

  • Individual pipelines per service
  • Docker containerization
  • Kubernetes manifests
  • Jenkins + GCP + GKE
  • GitOps ready (ArgoCD/Flux)
  • Distributed Tracing - Zipkin
Java 17Spring BootSpring CloudMySQLDockerKubernetesAuth0EurekaResilience4jJenkinsGCPGKE
99.2% Hit Rate10x Faster85% DB ReductionOPEN SOURCE

High-Performance Distributed Caching Solution

10x performance improvement with intelligent multi-level caching

Production-grade caching architecture demonstrating 10x performance improvement for high-traffic e-commerce applications. Implements multi-level caching strategy with Caffeine (L1 in-memory) and Redis (L2 distributed), featuring distributed locks via Redisson for cache stampede prevention, intelligent cache warming, and comprehensive monitoring with Prometheus and Grafana dashboards.

Multi-Level Cache Architecture

ClientHTTP RequestSpring Boot APIL1: Caffeine (In-Memory)• Ultra-fast: 1-5ms• Local to each instance• Size-based eviction (LRU)• No network latencyCache MissL2: Redis + Redisson• Distributed: 10-20ms• Shared across all instances• Distributed locks (stampede prevention)• TTL-based expirationCache MissMySQL Database• Persistent storage: 300-500ms• Source of truthMonitoring StackMicrometer(Metrics)Prometheus(Storage)Grafana(Visualization)1-5ms10-20ms300-500ms
ProductCacheService.javajava
1@Service
2@RequiredArgsConstructor
3public class ProductCacheService {
4
5 private final RedissonClient redissonClient;
6 private final Cache<Long, ProductDTO> caffeineCache;
7
8 @Cacheable(value = "products", key = "#id")
9 public ProductDTO getProductById(Long id) {
10 // L1: Check Caffeine (1-5ms)
11 ProductDTO cached = caffeineCache.getIfPresent(id);
12 if (cached != null) return cached;
13
14 // L2: Check Redis with distributed lock
15 RLock lock = redissonClient.getLock("lock:product:" + id);
16 try {
17 if (lock.tryLock(5, 10, TimeUnit.SECONDS)) {
18 // Only 1 thread queries DB (prevents stampede)
19 ProductDTO product = fetchFromDatabase(id);
20 caffeineCache.put(id, product);
21 return product;
22 }
23 } finally {
24 lock.unlock();
25 }
26 return fetchFromDatabase(id);
27 }
28}

Performance Improvements

Response Time
200ms to 15ms
Cache Hit Rate
99.2%
Database Load
-85%
Cache Stampedes
Zero

Caching Patterns

  • Multi-Level - L1 (1-5ms) + L2 (10-20ms)
  • Cache-Aside Pattern - Lazy loading
  • Distributed Locks - Redisson
  • Cache Warming - Predictive pre-loading
  • TTL Strategies - Data volatility based
  • Write-Through - Consistent updates

Monitoring & Observability

  • Micrometer - JVM and cache metrics
  • Prometheus - Time-series storage
  • Grafana - Real-time dashboards
  • Hit/Miss Tracking - Efficiency monitoring
  • Latency Percentiles - P50, P95, P99
  • DB Query Metrics - Before/after analytics
Spring Boot 3CaffeineRedisRedissonPostgreSQLMicrometerPrometheusGrafanaDocker
5 Payment ProvidersQR Check-inMulti-Tenant SaaSIN DEVELOPMENT

FitCI - Multi-Payment Gym Management SaaS

Comprehensive fitness platform built for the African market

Comprehensive SaaS platform for gym and fitness center management in West Africa. Features multi-payment provider integration supporting both international (Stripe) and local mobile money operators (Orange Money, MTN, Wave). Includes QR code-based member check-in, real-time analytics dashboard, automated billing, and staff scheduling—built specifically for the African fitness market.

FitCI Architecture (Excalidraw Style)

Next.js 15 FrontendTypeScript + Tailwind CSSQR Check-in + DashboardPort: 3000DeployVercelHostingAPI CallsSupabase BackendPostgreSQL + Auth + Real-timeRow Level Security (RLS)Real-time SubscriptionsPostgreSQLMembers, Payments, ClassesPayment ProvidersStripeCredit/Debit CardsMobile Money• Orange Money• MTN MoMo• Moov • WaveWebhooksKey Features✓ QR Code Member Check-in✓ Digital Membership Cards✓ Class Scheduling & Booking✓ Staff Scheduling & Payroll✓ Real-time Analytics Dashboard✓ CSV Data Import Tools✓ Revenue & Attendance ReportsLegend:BackendFrontendDatabaseExternal APIs

Platform Metrics

Payment Providers
5 Active
Transaction Success
98.5%
Real-time Updates
Supabase
Check-in Speed
<2s avg

Core Features

  • Member Management - Registration, profiles
  • QR Code Check-in - Contactless entry
  • Class Scheduling - Capacity management
  • Staff Management - Payroll, performance
  • Analytics Dashboard - Real-time metrics
  • Automated Billing - Invoice generation

Payment Integration

  • Stripe - International cards
  • Orange Money - West Africa leader
  • MTN Mobile Money - Pan-African
  • Moov Money - Regional wallet
  • Wave - Digital wallet transfers
  • Multi-Currency - CFA, EUR, USD
Next.js 15TypeScriptSupabasePostgreSQLStripe APIOrange MoneyMTN APIWave APITailwind CSSQR Code