Our goal is to build high-quality evaluation and training datasets that capture the real reasoning challenges lawyers and legal scholars face.