The Internet of Things (IoT) represents a transformative paradigm in modern mechanical engineering and industrial automation, enabling physical machinery to evolve into intelligent cyber-physical systems through continuous data exchange. The digitalization of traditional mechanical systems enables a transformation that, according to documented case studies, can reduce equipment downtime by 30-50% and achieve maintenance cost savings of up to 40%. This book chapter comprehensively addresses the conceptual framework and multi-layered architectural principles of the IoT ecosystem within the mechanical engineering domain, spanning device, edge, and cloud computing tiers. The primary focus of this study is the integration of Machine Learning (ML) algorithms with resource-constrained IoT devices and the extraction of actionable features from raw sensor data. In this context, the integration of signal processing techniques, such as Fast Fourier Transform (FFT) and wavelet analysis for vibration and acoustic signals, into ML pipelines is presented through novel architectural frameworks. By analyzing the applications of diverse ML paradigms on multi-modal data, the chapter thoroughly examines Edge AI, TinyML, and hierarchical sensor fusion architectures that overcome the limitations inherent to conventional cloud-centric approaches. Practical engineering solutions are exemplified through autonomous condition monitoring mechanisms deployed in remote scientific facilities with extreme environmental conditions, such as the Eastern Anatolia Observatory (DAG). Ultimately, by also discussing data privacy, federated learning, and 5G/6G infrastructures, this work provides a structured architectural guide demonstrating how IoT and ML integration transforms mechanical systems into autonomous decision-support systems.