Neural Graph Databases (https://www.ngdb.org/) represent a pivotal leap in structured reasoning and Neuro-Symbolic AI. By integrating logical query capabilities with in-context learning, NGDB addresses the structured reasoning limitations of current models, offering a robust Neuro-Symbolic framework that supports complex queries across structured modalities (e.g., graphs and tables). Potential research directions include data synthesis for structured datasets, expanding support for diverse model architectures, and developing new query types. This project is ideal for candidates with research experience in ML/NLP who are eager to build the "reasoning backbone" for future AI systems.