This paper presents an Intelligent Gesture Recognition System designed to facilitate seamless human-machine interaction within the Internet of Things (IoT) ecosystem. The system leverages Deep Learning to interpret complex hand gestures, translating them into control commands for an LED output. The architecture utilizes a high-resolution camera for image acquisition, integrated with a Python-based backend employing a Convolutional Neural Network (CNN) for real-time gesture classification.\\nData processing is handled via the Mediapipe and OpenCV libraries to extract hand landmarks, which are then fed into the trained model to ensure high accuracy and low latency. Upon successful recognition, control signals are transmitted via serial communication to an Arduino microcontroller, which serves as the hardware interface to toggle or dim the LED states.