Mahar Ibraheem

I designed an Edge-AI IoT security system for my bachelor's thesis, using PIR motion detection to trigger image capture and an on-device model to verify human presence before confirming an alarm.

I trained, quantised and converted a custom CNN to run under 1 MB on the ESP32-S3. The system performs local inference, and I measured its false-positive and false-negative rates against a PIR-only baseline.

In an industry collaboration with Axis Communications, I contributed to an Android application for Axis surveillance products, working on requirements, architecture, implementation, QA testing and documentation with an agile team.

I'm currently deepening my machine-learning knowledge at master's level and maintaining a Proxmox home lab. I also built a Python automated trading bot for MetaTrader 5 that runs on a Windows VM with a watchdog, autostart and Telegram alerts.