The next generation of wireless networks needs to support massive device deployments and ensure robust, widespread connectivity, particularly for edge devices in various Internet-of-Things (IoT) applications. To achieve this goal, two primary requirements should be addressed: (1) maintaining strong and reliable signal connectivity for a large number of devices, and (2) enabling real-time, on-device data processing for edge devices. To address signal connectivity, mmWave technology offers high throughput but suffers from limited propagation and high penetration loss, particularly in non-line-of-sight (NLOS) scenarios. This dissertation explores the use of reconfigurable intelligent surfaces (RISs) to optimize mmWave signal coverage. In particular, a stochastic geometry-based framework is utilized to evaluate the performance of mmWave communication assisted by multi-RIS links, considering both reflective and transmissive-reflective RISs. Additionally, multi-RIS deployment strategies are proposed to enhance indoor coverage by optimizing RIS placement using stochastic or full obstacle information. For on-device data processing, this dissertation investigates RIS-based over-the-air (OTA) computation, with the aim to reduce memory and computational demands by offloading computations from digital processors to the radio frequency (RF) domain. This dissertation investigates applications including a low-complexity direction-of-arrival (DoA) estimator leveraging RISs and RIS-based RF neural networks for machine learning inference in RF domain. Two RF network architectures are explored: fully connected layers with quantized RIS phase configurations and RF convolutional layers. Simulation results validate both applications, demonstrating the potential of RISs in enhancing wireless network efficiency.