Side-channel analysis holds immense potential for bolstering the security of products by identifying and mitigating vulnerabilities stemming from unintended information leakage. Traditionally, this approach necessitated extensive feature engineering to extract meaningful insights. However, the advent of machine learning has expedited the analysis process, albeit with varying degrees of success. While existing open-source models show promise, their performance often falls short of industry standards. In this project, we aim to develop machine learning-based model for side channel analysis, data pre-processing, critical cryptgraohic operation itentification and aligntment of traces by leveraging open-source datasets such as ASCAD as well as our lab equipment, and evaluate multiple state-of-the-art attention-based models to achieve satisfactory performance levels. To perform the tasks, we split the goals into 4 Thrusts.