Abstract: With the proliferation of digital threats in today's interconnected world, the need for efficient and effective automated malware detection systems has become paramount. Traditional signature-based methods often fail to keep pace with the evolving landscape of malware, necessitating the development of more sophisticated techniques. In this paper, we propose a robust approach to automated malware detection leveraging ensemble learning techniques. Ensemble learning, which combines the predictions of multiple base models, has demonstrated remarkable success in various domains, including cybersecurity. Our approach harnesses the diversity of ensemble methods to enhance the detection accuracy and robustness against adversarial attacks. By integrating diverse base classifiers such as decision trees, random forests, support vector machines, and neural networks, our ensemble model learns to effectively discriminate between benign and malicious software samples. Furthermore, we introduce novel feature engineering strategies tailored to capture the intricate characteristics of malware. These features encompass a wide range of attributes, including static file properties, dynamic behavioural patterns, and frequency-based representations. Leveraging these rich feature sets, our ensemble model can generalize well across different types of malware families and variants.