Do you use neural networks in your work? How exactly? What is your attitude towards them?

Yes, neural networks are actively used in work for solving various tasks: from natural language processing and image analysis to recommendations and forecasting.

For example, in a Go project, neural network models can be integrated through REST API calls to model services (such as TensorFlow Serving or PyTorch Serve) or by using libraries for working with ONNX models.

The attitude towards neural networks is rather positive, as they allow automating and improving the quality of decisions, but it is important to understand their limitations, the need for quality data, and the complexity of interpreting results.

Example of calling a model via HTTP from Go:

resp, err := http.Post("http://model-server/predict", "application/json", bytes.NewBuffer(jsonData))
if err != nil {
    log.Fatal(err)
}
// handle response

Thus, neural networks are a powerful tool that complement traditional development methods.