The field of sensing and machine learning is moving towards energy-efficient solutions, with a focus on developing innovative methods for human activity recognition, gesture recognition, and user authentication. Researchers are exploring the use of Wi-Fi Channel State Information (CSI) and other non-intrusive sensing methods to create privacy-preserving and contactless sensing approaches. Additionally, there is a growing interest in deploying machine learning models on resource-constrained devices, with a focus on developing compact and lightweight models that can operate independently of constant communication or external energy supply. Noteworthy papers include: HandPass, which presents a novel approach to biometric authentication using Wi-Fi CSI data for palm recognition, achieving an average F1-Score of 99.82%. Adaptive Forests For Classification, which proposes a novel approach that adaptively selects the weights of the underlying CART models, consistently outperforming Random Forests and Extreme Gradient Boosting in binary and multi-class classification problems. Enabling Vibration-Based Gesture Recognition, which proposes an energy-efficient solution deploying compact Neural Networks on low-power Field-Programmable Gate Arrays to enable real-time gesture recognition with competitive accuracy. IBIS, which introduces a novel hybrid architecture that integrates Inception-BiLSTM with a Support Vector Machine, achieving a movement recognition accuracy of nearly 99%. maxVSTAR, which presents a closed-loop, vision-guided model adaptation framework that autonomously mitigates domain shift for edge-deployed CSI sensing systems. STAR, which proposes an edge-AI-optimized framework that integrates a lightweight neural architecture, adaptive signal processing, and hardware-aware co-optimization to enable real-time, energy-efficient Human Activity Recognition on low-power embedded devices. ML-Based Preamble Collision Detection, which proposes a machine learning-based mechanism for early collision detection during the random access procedure, achieving over 98% balanced accuracy. Wireless Memory Approximation, which advocates two novel approaches for energy-efficient task-specific IoT data retrieval. Boosted Trees on a Diet, which presents a compression scheme for boosted decision trees, addressing the growing need for lightweight machine learning models.