At Hams.AI, I design and refine prompts for Arabic voice AI agents, turning client requirements into structured conversation workflows. I also organize client knowledge bases and test retrieval quality.
At SDAIA, I adapted SigLIP to more than 600K Makkah Al-Haram security-camera images, using image captioning, LoRA fine-tuning, and knowledge distillation. I also developed Swaher, a plain-text vehicle search tool for road CCTV footage.
I created a vehicle recognition pipeline and a real-time speech translation app during my SDAIA internship. The app used Whisper for transcription and Gemma for translation, with a FastAPI backend and React/Vite frontend.
My projects include an exact similarity search over Gaia DR3 star spectra and EyeVi, an assistive object-detection app for blind users. I graduated with a BSc in Artificial Intelligence from the University of Jeddah and completed the KAUST-Oxford AI Summer Program.
