Design and implement Edgey, a modern AI assistant as the intelligence layer for Clubedge. Goal: a permission-aware, context-driven AI that can reason, take actions, optimize workflows, and automate management tasks across the platform.

Responsibilities:

• Design the AI architecture (reasoning engine, memory, tools, orchestration)
• Implement the MCP action layer and integrate GPT models
• Build multi-tenant memory & RAG using Supabase pgvector
• Connect Edgey to backend modules with strict role & permission enforcement
• Develop automation workflows and document the full system

Technologies:

• GPT models (reasoning & function calling)
• MCP (Model Context Protocol) for AI-driven actions
• RAG with Supabase pgvector
• Node.js / TypeScript
• Supabase Postgres + Row-Level Security (RLS)
• Next.js (integration layer)

Notes: 1 trainee. Location not explicitly stated in the PDF; application is online.