We all have been amazed by successful applications of AI deep learning technology to areas such as image, speech, text, and games. However, we are also staring to recognize the weaknesses, including long training times and divergent issues, choice of architecture & hyperparameters, and complexity of the set-up. To address this, my recent work has involved the development of a entirely different paradigm in deep learning architecture to overcome these obstacles. Furthermore, I have applied my research to deep learning-driven Digital Twin Technology to engineering design, manufacturing, operational production optimization, and predictive maintenance. The conclusions from my work thus far have produced new comprehensive information/knowledge-based prognostics and health management e-tools for remaining useful life (RUL) predictions, predictive/proactive maintenance, scheduling optimization and synchronization high reliability, maintainability, and safety of plant equipment of all types of mass production or assembly processes. As more real and generative maintenance data becomes available and a deeper repository of such data is created, my deep learning-driven management will eventually take over the core competencies to form a new knowledge-centric proactive/predictive maintenance culture aiming to predict RUL o the components optimized operation quick real-time decision-making. I look forward to discussing the above work. Additionally, I will also briefly discuss my recent work in bridging deep learning and healthcare applications.