User profiling involves analyzing an individual's mindset based on their various activities. In today’s social media-driven world, users are engaged in a wide range of activities across different platforms. By examining these activities, any user’s sentiment could be assessed which could be utilized for various targeted applications as well. Typically, social media content consists of three main types: Textual posts, Images, and Videos. In this study, sentiment analysis focuses specifically on the videos shared by users on social media. The research is divided into two main phases. In Phase-1, deep learning techniques are employed to extract captions from the videos. In Phase-2, these captions are analyzed to determine the user’s sentiments by interpreting their contextual meanings. The experiment was conducted on a selection of user profiles from social media platforms such as Twitter (renamed as “X” from August 2023) and Facebook. The results and associated challenges are discussed in detail in this report.