The rapid proliferation of IoT devices has increased security and privacy vulnerabilities due to device resource restrictions and a lack of edge intelligence. To better understand how Supervised Machine Learning (ML) may be used at edge devices, this study examined how industry actors can use ML to improve IoT edge security. Despite the interest in ML for intrusion detection in IoT, edge device security is in demand as IoT devices spread. The current technique is computationally costly, and resource-limited IoT devices struggle to run ML algorithms. Using a mixed-method approach, this study uses EuX testbed and UNSW-NB 15 network datasets to train, assess, and finetune ML models for edge deployment. The study's findings present the model's performance, best features, compute time, and resource needs from an exploratory examination of the data sets. This study concludes that ML models can improve IoT real-time anomaly and intrusion detection by boosting edge device intelligence. However, ML deployments also require algorithm optimization and computational reduction.