Selected Works

Developed a deep learning CNN model for medical image classification that achieved near-perfect accuracy in detecting brain tumors from X-ray images. This project was completed during a month-long AI hackathon organized by Computiq and GIZ, working with a dataset of over 20,000 X-ray images.

Won first place in the "Sustainable IoT" hackathon among 9 competing teams. This project was developed as part of the CORTISSS program, a collaboration between Al-Nahrain University and Offenburg University. The success of this project earned me a DAAD Grant for study in Germany.

My final year project focused on early lung cancer prediction using machine learning. The model achieved 98.97% accuracy using Random Forest ensemble methods, with careful handling of imbalanced data using SMOTE (Synthetic Minority Over-sampling Technique).

A production-ready web application for embedding memories and stories directly into image metadata. Features include JWT authentication, end-to-end encryption, PWA capabilities, and full offline support.

An interactive landing page designed to teach Iraqi people no-code development, with a focus on prompt engineering and building AI-powered projects without writing code.

A real-time face recognition system integrated with conversational AI capabilities. The system recognizes users and engages in natural conversation using ChatGPT API and Amazon Polly for text-to-speech output.