From the last couple of years, the e-Content had greatly revolutionized the research and study prospects of higher education in universities. To enable the e-Learning and up-gradation of the e-Content, a dynamic cloud architecture using Machine Learning (ML) techniques is proposed in the paper. The idea here is to achieve twin goals. Firstly, allows content-based recommendation and authenticated use of e-Content for the knowledge and learning of the students and teachers. Secondly, allow up-gradation of the available e-Content to meet the latest knowledge standards. The e-Content is freely shared among the students and teachers from an educational point of view while allowing keyword search facilities to save time wasted due to random searching. However, to achieve the first goal, secure login and captcha verification are added to the designed cloud architecture. Further, Artificial Intelligence (AI) based architecture is integrated for the inclusion of more content to up-gradate the existing e-Content to support the latest knowledge base search of study material in universities. The simulation study comprises 100 simulation rounds for every variation in the number of neurons varied from 5 to 15. The performance of the proposed AI-based architecture is evaluated in terms of Mean Squared Error (MSE) and regression analysis to find the best neural structure in terms of neuron count. It is concluded that the proposed dynamic cloud architecture for e-Content recommendation demonstrates a minimum MSE of 0.0013 and a maximum R-value of 0.9566 when 9 neurons are used in the AI architecture.