Welcome to my Customer Orders App
The Customer Orders App is a production-grade, full-stack demo built end-to-end to showcase how modern systems are designed, delivered, and operated in the real world. This is not a tutorial project—it reflects the same architectural thinking, engineering rigor, and operational discipline I apply when building enterprise systems.
Built on .NET 9.0 and Nuxt 3, the application demonstrates how deterministic business logic and non-deterministic AI capabilities can coexist cleanly in a single system. It is fully containerized, continuously deployed via GitHub Actions, and runs live on a Linux VM behind a Caddy reverse proxy at rahullore.com.
From database migrations and API design to AI-assisted workflows and DevOps automation, every decision in this project mirrors real production constraints—performance, maintainability, observability, and ease of evolution.
- Production-Ready Backend Architecture – A complete .NET 9.0 API built using EF Core 9 (code-first), with clean domain boundaries and vertical slice architecture. All schema changes are managed through repeatable, versioned migrations—exactly how real systems evolve safely.
- End-to-End Containerized Deployment – The entire stack runs in Docker, with images published to Docker Hub and deployed onto a Linux VM. This mirrors real-world delivery pipelines used in cloud and on-prem environments.
- Modern DevOps & CI/CD Automation – Seamless deployments via GitHub Actions ensure every commit can be built, packaged, and released automatically, reducing human error and increasing delivery confidence.
- AI as a First-Class System Capability – The app demonstrates multiple AI patterns: chat completions, embeddings, and semantic search, integrated thoughtfully rather than bolted on—showing how AI augments systems instead of replacing core logic.
- Deterministic + Non-Deterministic Design Balance – Business-critical workflows remain deterministic and predictable, while AI is used where flexibility and intelligence add real value—reflecting responsible, production-grade AI adoption.
- SQL Copilot: Natural-Language Data Operations – A custom SQL Copilot enables querying, viewing, updating, and deleting data using plain English. This unlocks powerful capabilities for support teams and domain experts without deep SQL knowledge.
- High-Performance Full-Stack UI – Built entirely with Nuxt 3 and Tailwind CSS, the frontend is intentionally simple, fast, and responsive, focusing on usability and clarity rather than unnecessary complexity.
- Realistic Infrastructure Constraints – The system runs on a single Linux 22.04 VM with 8 GB RAM, hosting both the application and SQL Server—demonstrating practical resource planning and performance awareness.
- Security & Operational Readiness – HTTPS termination via Caddy, clean API boundaries, and controlled execution paths reflect real production hardening practices rather than demo shortcuts.
- Built by an Engineer Who Ships – This project reflects years of experience delivering real systems: thoughtful architecture, pragmatic trade-offs, operational awareness, and a strong focus on maintainability—not just “getting it to work.”