AI Training (RAG/NLP) – Graduation Project
Developed a banking-focused customer service assistant using Retrieval-Augmented Generation (RAG) for a graduation project. The role involved training models on Arabic text datasets to generate context-appropriate responses. Integrated and evaluated RAG systems for financial query handling in Arabic language contexts. • Focused on collecting and preparing Arabic text data for chatbot training. • Annotated and evaluated conversational AI outputs for accuracy and relevance. • Used TensorFlow and HuggingFace libraries to support model fine-tuning and evaluation. • Achieved 85%+ accuracy in response generation for banking-related use cases.