In this talk, we introduce a cutting-edge approach to detecting and mitigating interference in 4G/5G wireless networks using vision-based models. Our innovative solution leverages machine learning and computer vision techniques to detect interference in real time, addressing a significant challenge in the wireless industry. Our approach is unique in its ability to deploy these models on resource-constrained embedded systems within wireless radios, enabling practical implementation without compromising network performance. This technique not only enhances current 4G/5G systems but also lays the groundwork for integrated sensing and communication (ISAC) in future 6G networks. We will showcase the capabilities of this approach, demonstrating how it can revolutionize interference management and pave the way for more robust, efficient wireless communications.