Capstone Project: The AI Platform Portfolio

Overview

This is the culmination of your journey in the Platform Engineering & AI Infrastructure Course. You will build a complete, end-to-end platform that integrates an Internal Developer Platform (IDP) with a highly scalable AI Inference Engine. This project will serve as the centerpiece of your professional portfolio.

Requirements

Your final submission must include the following components:

1. Infrastructure as Code (Terraform)

• Complete Terraform configurations for provisioning the cloud infrastructure, including the Kubernetes cluster, networking (VPC), and IAM roles.
• The infrastructure must be modular and reusable.

2. Kubernetes Configuration (GitOps)

• A Git repository containing all Kubernetes manifests (Deployments, Services, Ingress, etc.).
• ArgoCD or Flux configured to manage the cluster state based on this repository.

3. Internal Developer Platform (IDP)

• A functional developer portal (e.g., Backstage) deployed to the cluster.
• At least one working “Golden Path” software template that allows a developer to scaffold a new AI-integrated microservice.

4. AI Inference Engine

• A production-ready deployment of an LLM serving engine (e.g., vLLM or Ollama) on a GPU node.
• Autoscaling configured based on queue length or token throughput.

5. Observability Stack

• Prometheus and Grafana deployed and configured to monitor the cluster and the AI inference engine.
• A custom dashboard displaying critical metrics (e.g., GPU utilization, request latency).

6. CI/CD Pipelines

• GitHub Actions (or similar) pipelines that automatically build and test the infrastructure and application code upon commit.

Deliverables

• A link to your public GitHub repository containing all the code.
• A comprehensive README architecture document detailing your design decisions, trade-offs, and instructions for deploying the platform.
• A 5-10 minute video walkthrough demonstrating the platform in action.