الذكاء الاصطناعي والأتمتة (AI / Automation)
AI-Powered Pharmacy Assistant Telegram Bot
An AI-powered pharmacy assistant Telegram bot built with advanced RAG automation workflows. The system integrates cloud-hosted n8n workflows and Qdrant vector retrieval for accurate medical data querying, alongside Google Sheets synchronization. It includes robust LLM fallback logic to ensure consistent and reliable responses. Presented for An-Najah National University in collaboration with UNDP and the Women's Affairs Center, the system was tested using Docker and is fully scalable and deployable to any cloud environment.
Overview:
Designed to provide a reliable AI-powered pharmacy assistant on Telegram, this system leverages a Retrieval-Augmented Generation (RAG) architecture to deliver fast and accurate responses to medical inquiries.
Technical Challenge:
The objective was to build an automated system capable of retrieving complex medical data with high precision from dedicated sources, bypassing standard generative limitations, while ensuring the data could be easily managed and updated by administrators.
The Solution:
The architecture utilizes cloud-hosted n8n workflows (rather than local instances) to connect a Telegram interface with a cloud Qdrant vector database for precise querying. To streamline data management, the system synchronizes seamlessly with Google Sheets. Furthermore, robust LLM Fallback Logic was implemented to guarantee continuous service availability in case a primary model fails.
Reliability & Scalability:
The system was successfully tested within a Docker environment and engineered to be fully scalable and deployable to any cloud environment.
The Outcome:
The project was successfully presented as an applied intelligent automation model for An-Najah National University, in collaboration with UNDP and the Women's Affairs Center.
Developed by Engineer Sami Meqdad