The main achievement of this project is the ability to start from a complex, unstructured description of a research problem written in natural human language and convert it into an explainable recommendation process. Rather than performing only keyword or semantic matching, the system interprets the problem, decomposes it into relevant dimensions, identifies matching researchers and grants, and provides reasoned justification — including strengths, limitations, and relevance assessment. In this sense, the system moves beyond retrieval toward AI-supported academic reasoning.