The increasing adoption of wireless network related applications leads to an increasing number of users seeking to access scarce radio spectrum resources. The use of multiple access techniques enables the realization of this goal. An important multiple access technique in this regard is the non-orthogonal multiple access (NOMA) method. In NOMA, it is important that receivers can execute multi-user detection (MUD). In this regard, the use of artificial neural networks (ANNs) is beneficial. This is because the ANN can handle the challenge of non-linear-multi-user detection. The challenge arises when a significant number of users seek to access the spectrum in a manner that the use of closed-form expression related methods is insufficient. The research presents the design and performance of an ANN-based multi-user detector, performance shows that the training, testing, and validation mean square error is 0.023103 at epoch 2.