The objective of the proposed research is to demonstrate the use of electromagnetic (EM) sensing as a tool for inferring information from biological and embedded systems. EM sensing can be a powerful tool to acquire information about systems that are otherwise hard to access. In this proposed thesis, we apply EM sensing to both biological and embedded systems. We first investigate EM modeling and measurements to track plant water content. Conventional methods to determine the relative water content of a leaf require the destruction of said leaf. In contrast, we develop a simple EM model and demonstrate that EM transmission loss in the Ka-band strongly correlates with the leaf water content. To overcome the practical limitations of transmission measurements, we develop a model and measurement technique with a single antenna that tracks the loss due to leaf water content via EM backscattering. We then utilize EM side-channels to profile embedded systems. We use a pair of microcontroller units (MCUs) that exchange messages using different serial protocols. Measuring the magnetic near-field, we are able to distinguish different boards, protocols, data rates, and even message contents with high accuracy. We also demonstrate that this EM side-channel radiates, and we are able to perform similar classification using far-field data recorded at lower sampling rates and in a noisy lab environment. Finally, we propose to emulate a CubeSat platform using a single-board computer (SBC) with a sensor extension board. Given the demonstrated feasibility to distinguish serial communications, we aim to demonstrate that periodic sensor polling and communication patterns can be profiled using EM side-channels. This approach would serve as a foundation for profiling inaccessible embedded systems.