A neural network is a mathematical model of many interconnected nodes ("neurons") arranged in layers. An input is sent through the network, transformed layer by layer and finally an output is produced. During training, the weights between neurons are adjusted so the network solves its task better and better.
Large networks with millions or billions of parameters are called "deep" – hence the link to deep learning. Specialised architectures such as convolutional neural networks (CNN) for images, recurrent neural networks (RNN) for sequences and transformers for language have driven the AI leaps of recent years.
In HR and recruiting, neural networks underpin modern tools for CV analysis, job matching and generation. They are powerful but often hard to explain – a key reason for demands for transparency and human final decisions.
Lunigi uses neural networks in the background but combines them with clear filter rules and human feedback in the email workflow.