{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/fd28447569544670a5b14ce2f3c4e8c0\" frameborder=\"0\" width=\"1092\" height=\"819\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":819,"width":1092,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":819,"thumbnail_width":1092,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/fd28447569544670a5b14ce2f3c4e8c0-697fd1e256880e69.gif","duration":253.567,"title":"Elevant Digital AI Learning Assistant (Discord + RAG System)","description":"I developed an AI-powered learning assistant for Elevant Digital designed to support students and internal academic teams by providing instant answers from learning materials.\n\nThe system is built as a full Retrieval-Augmented Generation (RAG) pipeline and is integrated with Discord for real-time interaction.\n\nProject Overview\n\nThe workflow is designed around two production-grade API endpoints:\n\n/ingest endpoint for processing and indexing learning materials\n/ask endpoint for handling user queries and generating responses\nSystem Architecture\nLearning materials (PDFs) are uploaded via Google Drive\nA Make.com workflow triggers the ingestion pipeline automatically\nDocuments are processed, chunked, and converted into embeddings\nEmbeddings are stored in Qdrant vector database for semantic search\nA Discord bot handles student queries in real time\nRelevant context is retrieved from Qdrant and passed to Gemini AI\nThe model generates accurate, context-aware answers\nThe system is deployed on Render\nImpact\n\nThe system enables students to:\n\nAsk questions directly from Discord\nReceive instant explanations from course materials\nReduce dependency on immediate teacher availability\nAccess structured academic knowledge in real time"}