What is NumPy's internal ndarray object architecture?

ndarray is a C-backed structure containing a data buffer, shape, strides, and metadata.

NumPy's ndarray is not a Python container but a thin Python wrapper around a contiguous C memory block. It stores metadata such as shape, strides, dtype, and pointer to raw data. This separation allows NumPy to perform vectorized operations without Python-level loops. The design is optimized for cache efficiency, low-level memory access, and interoperability with C/Fortran libraries.