This work presents a novel approach to the diagnosis of lung cancer by greatly improving diagnostic accuracy by utilizing the power of You Only Look Once Neural Architecture Search (YOLO-NAS). Through the combination of cutting-edge neural network search algorithms and the state-of-the-art object identification system YOLO-NAS, we have created a highly accurate and efficient model for recognizing lung cancer from medical imaging. The system’s performance was thoroughly assessed using an extensive lung scan dataset. At an intersection over the union (IoU) threshold of 0.50, the system obtained an impressive mean average precision (mAP) of 0.968. This high degree of precision shows how well the model can identify lung cancer, outperforming other techniques in terms of speed and accuracy. This approach not only provides a scalable methodology for improving diagnostic processes across a range of medical illnesses, but it also makes lung cancer detection more accurate and timelier. The findings have important ramifications for medical diagnostics, pointing to a new era in which AI-driven techniques will be vital in the identification and treatment of diseases.