This white paper explores how the AI Engine graph programming model is defined based on the Kahn process network (KPN). The KPN model helps to make the data flow parallel, which improves the overall performance of the system. Programming the AI Engine array requires a thorough understanding of the algorithm to be implemented, the capabilities of the AI Engines, and the overall data flow between individual functional units. AI Engine kernels are functions that run on an AI Engine and form the fundamental building blocks of a data flow graph specification. The data flow graph is a KPN with deterministic behavior. This white paper also includes an example design to illustrate a data flow graph with four AI Engine kernels that form the fundamental building blocks of a data flow graph specification. This example also demonstrates a data flow stall in the design and provides a solution.