SimLottery MCP

Built with

• Idea: OpenAI ChatGPT (GPT-5.2)
• Graphic Design: TBD
• Requirements: PROMPT.MD

Architecture Design

Designed a staged system with clear separation of concerns:

• Backend services for simulation + persistence
• A small public API (BFF) used by the dashboard
• A read-only web dashboard for public visibility

Implemented AWS infrastructure as code using AWS CDK, deploying:

• Containerized compute (ECS Fargate)
• A managed database (Aurora PostgreSQL)
• Static website hosting (S3 + CloudFront)
• A public API surface (API Gateway)

Development

Worked iteratively in small, reviewable changes with frequent validation.

AI-assisted engineering was used to speed up implementation and troubleshooting:

• Tooling: Cline (VS Code) with an LLM backend (Claude via API).
• Typical use: code navigation, refactors, patch generation, and deployment orchestration.

Testing

Performed local and production smoke checks:

• Built the web app and validated pages render correctly.
• Verified the Historical Matches page refreshes and shows newly available match results.
• Validated deployed assets via CloudFront and confirmed navigation in a real browser.

Deployment

Deployment was automated through infrastructure as code (AWS CDK).

Web releases followed a repeatable flow:

• Build static assets (npm run build)
• Deploy via CDK (upload + CloudFront distribution update)
• Invalidate CloudFront cache to ensure customers receive the newest version.

AI inside

None

Tech Stack

AI recommended tech stack