This research introduces an AutoML-based framework for optimization of intrusion detection pipelines for security of networked system. With AutoML, the framework can automatically scan through the process of picking and tuning machine learning models to identify or prevent potential intrusions. This approach hinges on the use of XGBoost, a powerful and scalable machine learning algorithm with better classification accuracy, but also more robust performance under various network conditions. To this end, we propose a solution that is streamlined and adaptive to intrusion detection while coping with large-scale datasets and evolving security threats. The framework optimizes the automation of intrusion detection system by minimizing the need for human intervention while improving the scalability of intrusion detection systems in modern complex network environments.