Anti-doping analysis is vital for combating cheating in sports, with the Athlete Biological Passport monitoring haematological and steroidal markers. WADA's (World Anti-Doping Agency) recent findings highlight the risk of urine sample swapping, where athletes substitute their sample with a clean one to evade detection. To address this, DNA analysis is used to confirm suspicious cases, guided by factors like sport demands, past violations, and athlete intelligence. However, identifying swapped samples requires an adaptive model to flag mismatched samples based on steroid profile similarities. This project aims to develop an AI-driven pattern recognition tool to analyze steroid profiles, identify key indicators, and deploy advanced deep learning algorithms, creating a robust pipeline to detect sample substitution efficiently.