The COVID-19 times was one of the most revolutionary times that made many changes in education societysuchastransferring the learning process from in-person to virtual learning environment. The virtual learning environment camewithits own advantages and disadvantages. Advantages such as remote access of the lectures, consistent learning etc, but it camewith tremendous disadvantages such as lack of self-discipline, social isolation and limited interaction which made everyonerealize the importance of teachers in-person interaction and learning. This research gives the best solution available byusingbagging (bootstrap aggregating) ensemble learning with 1D convolutional neural networks (1D CNN), 1Dresidual networks(1D ResNet), and hybrid deep learning models. The model uses bagging which helps learning to connect with CNNmodel andanother the ResNet model to evaluate scores on students’ engagement detection model. The combination of CNNandResNethelps the model to make more accurate and stable predictions. This research brings out the solution using DAiSEE(Dataset forAffective States in E-Environments) which monitors the student engagement during the virtual class environment bymonitoring facial expressions such as frustration, boredom, lack of concentration etc. Employing these modernmodelsenhances the accuracy and yielding capacity of the model compared to individual deep learning models and hence helpsthemodel to make most accurate predictions. By the development of this model, it enables a proper learning environment whichdemonstrates an evolution in the modern learning process. Keywords: Bagging, 1D Convolutional Neural Networks (1D CNN), 1D residual networks (1D ResNet), DeepLearning, DAiSEE (Dataset for Affective States in E-Environments), Virtual learning, Engagement detection model