This article introduces an AI-driven smart home project developed by Intel engineer Oliver Hamilton using the ​LattePanda Mu development board. By integrating ​computer vision and ​edge computing technologies, the system automates home lighting control through human pose recognition. The project leverages ​Intel Geti for model training, ​OpenVINO for low-power inference on the ARM-based LattePanda Mu, and ​MQTT combined with ​Node-RED to orchestrate smart lighting (Lifx/Hue) based on user status (e.g., "at desk" or "away"). The solution addresses key smart home challenges by minimizing manual intervention and optimizing energy consumption, demonstrating the potential of compact, high-performance hardware in creating responsive AI-driven living spaces