Chaimae ELFATHI

Collaborated in the design and implementation of a cutting-edge AI system, integrating Azure OpenAI and CrewAI for a multi-agent framework to enhance decision-making processes. Contributed to optimizing workflows and improving overall system efficiency with context-aware AI techniques.

Designed and developed a dual-agent AI system to enhance autonomous driving assistance using large language models (LLMs), computer vision, and retrieval-augmented generation (RAG) for real-time decision-making. Key Contributions: - Developed a Perception Agent utilizing an image-to-text pipeline to process visual data and generate structured textual descriptions. - Built an Assistance Agent incorporating Visual Question Answering (VQA) and fine-tuned Llama 3 models for intelligent, context-aware driving recommendations. - Processed 10,000+ images and 5,000+ QA pairs from the NuScenes dataset to improve model accuracy and real-world applicability. - Integrated RAG-based retrieval to enhance emergency response efficiency, achieving a 30% improvement in decision accuracy. - Applied deep learning, NLP, and multimodal AI techniques to create a robust system for autonomous vehicle support. + This project demonstrates expertise in computer vision, natural language processing (NLP), LLM fine-tuning, multimodal AI, autonomous systems, and real-time decision-making for autonomous systems.