The paper focuses on the problem of technical social engineering attacks that encompass the manipulation of individuals to reveal sensitive information, execute actions, or breach security systems. These exploits frequently capitalize on human psychology, trust, and a lack of vigilance to attain unauthorized entry to networks, systems, or data. In contrast to traditional social engineering tactics that center on psychological manipulation, technical social engineering attacks employ technological means and strategies to manipulate and beguile individuals. The paper presents an attempt to detect social engineering attacks. The approach utilized four machine learning algorithms (decision tree, random forest, K-nearest neighbor, and extreme gradient boosting). The analysis is centered on data collected from network hosts, which may serve as indicators of a potential social engineering attack. The empirical results demonstrated high detection accuracy.