In an era marked by escalating ransomware attacks,conventional detection methods often prove insufficient, leaving individuals and organizations susceptible to substantial data loss and operational disruption. This talk presents a multi-layered defense system utilizing eBPF for real-time monitoring. Machine learning algorithms are then employed to identify patterns indicative of ransomware, enhancing threat detection capabilities. Additionally, honeypots are deployed to validate ransomware presence during encryption, allowing for targeted response actions. By integrating these technologies into a cohesive framework, the system aims to comprehensively protect against ransomware by identifying and neutralizing threats at different stages, thus improving overall security resilience and understanding of ransomware behaviors.