On Fahes, my graduation project now being developed in collaboration with the Public Security Directorate, I developed the complete AI system for vehicle inspection. It uses YOLO11 to assess body damage, tires, lights and window tint.
I fine-tuned Qwen2.5-VL to analyze vehicle images, describe damage, review YOLO11 outputs and generate inspection reports. In measured project tests, the system completed visual inspection, analysis and report generation in under one minute.
I independently built a bilingual Arabic and English research agent using LangChain and LangGraph. It uses RAG over web sources and MCP integration to turn user-supplied topics into research reports with source references.
My other projects include comparing CRF and Transformer-based NER models on CoNLL-2003, where the CRF model achieved an F1 score of 0.80, and building a real-time clothing and color recognition system. I also led scripting, scene design, AI generation, editing and delivery of Fahes training and awareness videos.
