Ridgemont High

Production-grade AI education platform that teaches artificial intelligence through 80s/90s-inspired culture. Full Turborepo monorepo with 12 packages: auth (NextAuth v5, JWT, 6-role RBAC), config (Zod-validated env, IP_MODE, feature flags), content (seed lessons for 3 tracks), db (Prisma, 25+ models, pgvector), labs (7 lab engines with session management), mcp (tool registry with permission gateway), news (curation engine with Then vs Now briefings), personas (10 AI teacher definitions with IP mode resolver), rag (chunking, embedding, hybrid retrieval, RRF ranking, citation formatting), security (5-layer moderation pipeline — rate limiting, prompt injection detection, blocked content filter, age guardrails, output moderation with audit logging), types (250+ shared interfaces), and ui (design system with retro effects — scanlines, neon borders, VHS dividers, arcade badges). 10 AI teacher personas: Professor Hartwell (AI Foundations), Cipher (AI Coding), Ms. Chen (Prompt Engineering), Coach Williams (Ethics & Society), Dr. Nakamura (Agent Frameworks), Ms. Rivera (Creative AI), Mr. Kowalski (RAG & Search), Principal Torres (Platform Guide), Ms. Dawson (Current Events), DJ Lex (Creative AI). 9 learning tracks from beginner to advanced: AI Foundations (12 lessons), Prompt Engineering (10), Ethics & Society (10), Agent Frameworks (12), RAG & Search (10), MCP & Tools (8), Creative AI (10), Deep Learning (14), AI & Business (8). 7 interactive labs: Prompt Lab, Agent Lab, RAG Lab, MCP Lab, Code Lab, Media Analysis, Project Studio. Full admin panel with content studio, user management, moderation queue, analytics, and news curation. IP Mode system for legal safety on cultural references (inspired vs homage). PostgreSQL 16 + pgvector for semantic search, Redis 7 for sessions and rate limiting, MinIO S3 for media storage. 58 passing tests. Docker Compose orchestration. Vercel deployment target.