KAFL | NEXUS
AI-powered education analytics platform that converts exam result PDFs into actionable student guidance.
- Record
- Project 020
- Documentation
- Technical notes included
KAFL | NEXUS is an AI-assisted academic performance and strategy platform for high school students preparing for TYT/AYT exams. The system analyzes trial exam results at student, classroom, and school level, produces trend charts, and turns raw exam data into personalized guidance reports. It includes PDF result ingestion, structured PostgreSQL storage with Drizzle ORM, student and exam management, roster sync, student merge tools, API key vault controls, prompt presets, and a queue-based AI analysis flow. The interface uses a dark-mode-first Nebula design system with glassmorphism panels, neon accents, animated transitions, responsive components, and administrator workflows built for daily school operations.
System components
- Next.js App Router
- React Server Components
- TypeScript
- PostgreSQL / Neon
- Drizzle ORM
- Zod Validation
- PDF Ingestion
- AI Provider Layer
- Admin Dashboard
Data under tension
PDF results -> JSON extraction -> Zod validation -> DB upsert -> Queued AI analysis -> Student and class dashboards
Tradeoffs
- Maintaining reliable PDF extraction across changing formats
- Protecting provider API keys and student data
- Keeping AI feedback actionable and auditable