In this tutorial, we will introduce how machine learning can be utilized to solve the control and game problems motivated by the applications in finance. We will first start with introducing the stochastic optimal control problems with only 1 agent (such as a trader, employee, or a bank) and discuss how deep learning can be used to solve these problems. Later, we will move to the case where we have large number of agents. These problems are specifically important for financial applications to model the decision making of interacting traders or financial institutions. We will look at different setups such as models with cooperative agents, or non-cooperative agents and discuss different ways to solve these problems with machine learning. Later, we will discuss different extensions such as introducing multiple populations, network interactions among agents, or a significant player into our model and discuss how the methods can be extended to these situations.