Personalized content recommendations are essential for improving user experience in the current digital era. A recommendation system based on a hybrid deep learning and machine learning approach is presented in this paper. It uses facial expression analysis to recommend movies or music. To identify and categorize a user's facial expression into one of four moods— happy, sad, angry, or neutral—the system uses the deep learning model ResNet-50. The user is presented with music and movie recommendations based on the mood that was detected. For music, the system provides songs by genre (love, melody, folk, sad, friendship) and language (Hindi, English, Telugu). It offers movie recommendations according to specific genres (comedy, action, horror, and love). The recommendation engine optimizes music and movie recommendations by utilizing Linear Regression and Logistic Regression. This system offers an intuitive, emotion-driven approach to entertainment recommendations, bridging the gap between artificial intelligence and user preferences.