This research investigates intelligent agents capable of perceiving, reasoning, and acting within the physical world through direct interaction. We focus on Embodied AI, where intelligence emerges and is shaped by the agent's physical form and its sensory experiences within dynamic environments. Our work involves leveraging robotic platforms, such as robot arms and mobile systems equipped with various sensors, to gather real-world data and develop algorithms that enable skillful manipulation, navigation, and adaptation. Key goals include advancing robotic capabilities in complex, unstructured settings like manufacturing pipelines and fostering more intuitive human-robot collaboration.