Pneumothorax detection and segmentation in chest X-rays is a critical task in medical imaging that requires accurate and efficient diagnostic tools. This paper addresses the challenge of balancing computational efficiency with high accuracy in pneumothorax detection and segmentation, particularly in the context of limited computational resources and the need for real-time analysis. Existing solutions often rely on single-scale models or complex architectures that may struggle with generalization across diverse datasets or require significant computational power. To overcome these limitations, I have built a novel multi-scale ensemble approach that combines U-Net models trained on different image resolutions, leveraging a ResNet34 encoder and incorporating advanced techniques such as spatial and channel squeeze-and-excitation (scSE) modules. My Multi-Scale-Unet-Ensemble model has achieved a dice similarity coefficient of 0.855. I've used the SIIM ACR Pneumothorax Segmentation dataset, which is larger and of higher quality, ensuring improved performance and robustness across diverse cases.