Project 07: Production Cloud AI Platform

Objective

Design and document a production-style cloud AI platform with secure networking, compute, data stores, AI services, observability, backup, cost estimation, CI/CD, and Infrastructure as Code.

Reference Architecture

Internet -> CDN/WAF -> Load Balancer -> API Gateway -> AI Service + Backend API -> PostgreSQL/Redis -> Vector Database -> Object Storage -> External or Hosted Model

Tasks

• Draw architecture diagram.
• Define network boundaries.
• Define IAM and RBAC.
• Define database and cache strategy.
• Define vector database strategy.
• Add secrets management plan.
• Add logging and monitoring plan.
• Add autoscaling plan.
• Add backup and disaster recovery plan.
• Add cost estimate.
• Add CI/CD pipeline plan.
• Add Terraform module plan.

Deliverables

• docs/architecture.md
• docs/security.md
• docs/operations.md
• docs/cost-estimate.md
• docs/disaster-recovery.md
• infrastructure/ IaC skeleton
• ADRs for major decisions

Tips

• Design for failure explicitly.
• Document which services are managed and which are self-hosted.
• Include model inference cost in the total cost estimate.
• Do not ignore logs, metrics, and traces.