Artificial Intelligence (AI) is becoming the striking force of this era. Machine Learning (ML), deep learning (DL), and reinforcement learning (RL) are the three methods to train AI-based systems to elicit intelligence from row data. All these three methods need massive computation power to train AI tasks. CPUs can handle normal ML, DL, or RL tasks, but normally, it is difficult to handle complex tasks. Especially for any DL tasks, it needs a Graphics Processing Unit (GPU) or Tensor Processing Unit (TPU) to accomplish the AI tasks. GPU and TPU are costly and inconvenient for all AI enthusiasts or developers; sometimes, it might hinder their research and development process. This research aims to pave the way for optimally handling ML and DL tasks in the CPU. One complementary method to convey AI tasks besides personal GPU is free or paid cloud systems like Google Colab, AWS, IBM, or Azure Cloud. This option might not ensure data security and confidentiality, and they are sometimes insufficient with respect to the need. However, the proposed method provides data privacy and integrity compared to the typical cloud-based system. After implementing the developed algorithm for calculating latency and speed, we found the lowest training time for 16 CPU cores (DL:40s, ML: 5s) and the highest training time for 1 CPU core (DL:58s, ML:58s). This approach can be the best reference for deploying models to the end systems.