Embodied AI aims to create intelligent agents capable of interacting with and reasoning about the physical world. While deep learning has enabled impressive advances in perception and control, purely neural approaches often struggle with generalization, interpretability, and reasoning. Neuro-symbolic approaches offer a promising alternative by integrating symbolic reasoning and planning with neural representations, combining the strengths of both paradigms. The Ph.D will develop new neuro-symbolic methods for Embodied AI, focusing on how these hybrid approaches enhance decision-making in dynamic environments. The Ph.D will develop learning and planning algorithms, demonstrating improved performance on embodied tasks such as navigation, manipulation, and human-robot interaction.