EGS AI

When AI chat tools first emerged, students across Egypt turned to them for help with their schoolwork. However, a major problem quickly surfaced: standard AI models would provide scientifically correct answers, but they were completely outside the official Egyptian Ministry of Education curriculum. Students studying for secondary and middle school exams found themselves confused by university-level theories or foreign terminology that clashed directly with ministry guidelines and model answers, leading to declining grades and deep frustration. This sparked the idea for EGS AI. The core vision was simple: why should a student study a structured national curriculum only to receive chaotic answers from the open internet? To solve this, we engineered a powerful Hybrid Retrieval-Augmented Generation (RAG) engine. Instead of letting the AI guess, the system ingests official textbook PDFs page by page, structuring them into exact units and lessons. When a student asks a question, the engine searches exclusively within their specific textbook chunks—combining dense vector search, lexical BM25, and cross-encoder reranking—to deliver precise, curriculum-bound answers. What started as a response to student complaints evolved into a complete, intelligent educational platform tailored specifically for Egyptian students, ensuring that every explanation, exam, and flashcard aligns strictly with what they actually need to succeed in their official exams.