The extended Berkeley Packet Filter is a technology that allows the execution of sandboxed programs in the Linux kernel, with a high potential for packet filtering. The eXpress Data Path (XDP) further extends eBPF and allows high-performance packet processing on network interface cards. eBPF and XDP are used to implement intelligent security solutions powered by ML models to detect anomalous behaviour. eBPF-LAIN is herein proposed as an intrusion detection system (IDS) with XDP offload support. eBPF-LAIN is capable of detecting port scanning using a random forest model, surpassing the performance of traditional Intrusion Detection Systems like snort and other solutions. The evaluation results demonstrate that eBPF-LAIN is able to process a higher number of packets with minimal overhead. The results achieved demonstrate that eBPF-LAIN can be used to develop a solution that provides promising detection capabilities and performance, suitable for real-world scenarios.