Modern information technologies and the advent of machines powered by Artificial Intelligence (AI) have strongly influenced the world. The industry must evaluate what it will take to support AI as it moves towards mainstream. The rapid evolution of algorithms, software, and hardware to support AI requires a framework to understand the challenges and create a context to meet those challenges. In recent times there has been increase in usage of AI based applications built using Tensor Flow package, which is a widely used library, by developers to build deep learning applications, it is necessary to construct a performance engineering methodology and approach to test and optimize deep learning applications using tensor flow package. This paper includes the performance engineering approach and performance optimization of Deep learning applications using Tensor Flow Package.