Projects

Selected builds that show how I design, ship, and operate software across AWS and Google Cloud—including product delivery, AI-assisted investigation workflows, and reliability engineering.

AI-Assisted Incident Platform

IncidentLens AI

An AI-assisted incident detection and investigation platform that turns application errors into structured incidents, persists them on AWS, then enriches them with Amazon Bedrock for advisory investigation support in a React operator UI.

ReactTypeScriptNode.jsAWSAmazon BedrockTerraformGitHub Actions

Event-driven path: CloudWatch Logs → processor Lambda → DynamoDB persistence first, then Bedrock enrichment
Bedrock returns an AI-assisted summary, possible cause, and recommended investigation actions as advisory aids
React incident console with open → investigating → resolved lifecycle; SNS for eligible high/critical incidents
Terraform IaC with GitHub Actions OIDC deploy—no long-lived AWS keys in CI

Live Demo GitHub

Production Full-Stack Project

PitchPulse26

A production World Cup prediction platform built with React, Node.js, PostgreSQL, and serverless AWS infrastructure. Used by 20+ registered users with 600+ predictions during a live tournament.

ReactNode.jsPostgreSQLAWSTerraformGitHub Actions

Full-stack product covering authentication, match predictions, standings, and leaderboards
Express API on AWS Lambda behind API Gateway, with Prisma and Neon PostgreSQL
Infrastructure managed with Terraform; deployments automated through GitHub Actions
CloudWatch dashboards, alarms, and a documented rollback path

Live Site Case Study GitHub

Cloud Reliability Project

ApplyTrack API

A FastAPI backend demonstrating Python APIs, Docker packaging, a historical Google Kubernetes Engine deployment, CI/CD with Cloud Build, and practical observability with Prometheus metrics and local Grafana.

PythonFastAPIDockerGoogle CloudKubernetesPrometheusGrafana

FastAPI + Pydantic + SQLAlchemy + Alembic with PostgreSQL persistence
Historical GKE deployment via Cloud Build → container image → Deployment + LoadBalancer Service (cluster later deleted for cost control)
Kubernetes livenessProbe on /health and readinessProbe on /ready with a database connectivity check
Prometheus metrics exposed by the API; Prometheus/Grafana validated locally with Docker Compose and documented SLI/SLO analysis

GitHub