This project applies supervised machine learning to assess the pathogenicity of single nucleotide and structural genetic variants associated with neurological disorders. Using models such as RidgeCV and HuberRegressor trained on data from Ensembl, ClinVar, and gnomAD, the system generates risk scores and classifications. High model precision was achieved through cross-validation. An accompanying app translates these models into a practical tool for genetic analysis, early diagnosis, and clinical research.