Rust-based inference engines and local runtimes have appeared with the shared goal: running models faster, safer and closer ...