Developed a foundational data structures library for the postechDNN/postechDNN repository, focusing on enabling efficient data processing and scalable performance within deep neural network workflows. The work involved implementing core data structures such as AVLTree, BBST, DCEL, RBTree, SegL1Voronoi, SegTree, and a compressed quadtree, all designed to support fast and reliable graph and geometry operations. Using C++ and leveraging algorithmic design principles, the developer established a reusable, high-performance core to underpin future features and optimizations. This initial contribution laid the groundwork for performance-critical components, emphasizing maintainability and extensibility for subsequent development in DNN pipelines.