Reservoir computing is a computational paradigm within the domain of recurrent neural networks, characterized by a fixed, non-trainable hidden layer (the "reservoir") that maps sequential input data into a high-dimensional state space. This framework investigates efficient methods for processing time series data and modeling dynamical systems by leveraging the complex, non-linear dynamics and memory properties inherent in the reservoir's transient response, with only a simple output layer being trained. Its significance lies in offering competitive performance on many sequential data tasks with substantially reduced training complexity compared to traditional recurrent network architectures.