Machine learning has achieved significant progress in recent years, with algorithms matching human capabilities in numerous tasks. However, the main hurdle lies not just in creating these models, but in utilizing them optimally in real-world applications. This is where inference in AI becomes crucia