Modern C++ offers powerful abstractions for high-performance computing, but achieving fully SIMD-generic code remains a challenge. This work presents a practical approach to explicit SIMD vectorization using modern C++ features, enabling unified, portable, and efficient code generation for both vectorized and scalar (residual) loop iterations. By encoding SIMD type information directly within the loop index via a generic lambda interface, the proposed simd_access::loop abstraction allows a single source code path to handle both SIMD and scalar execution seamlessly. The solution supplements the upcoming std::simd library (C++26).