Developed a new CPU-targeted SVD algorithm option for the ROCm/jax repository, expanding JAX’s svd functionality to support a QR-based approach alongside the existing divide-and-conquer method. The work involved implementing a Foreign Function Interface wrapper in C++ and Python to connect with LAPACK’s gesvd routine, enabling QR-based SVD computations on CPU backends. This addition provided users with an alternative algorithm path, allowing for improved numerical properties and performance trade-offs. The API was extended to let users select the QR method explicitly, demonstrating depth in linear algebra, numerical methods, and cross-language integration within a high-performance scientific computing context.