This chapter explores the architecture of feedforward neural networks, key structures in AI that emulate human brain functions. We illustrate their hierarchy, from input to output, highlighting the roles of weights, biases, and activation functions. Central to learning in these networks is the backpropagation algorithm, which iteratively improves predictions through calculus-based adjustments. While powerful in pattern recognition, feedforward networks have limitations in reasoning tasks, setting the stage for the Neuro-Symbolic AI framework. This chapter provides a foundational understanding of feedforward networks’ role in bridging data-driven perception and logic-driven cognition, preparing readers for integrating neural and symbolic AI methodologies.