Managed Rust-GPU SDK for deploying custom inference kernels on consumer-grade hardware.

Dev teams lack safe, low-level GPU access for custom ML acceleration without C++ complexity.