Genetic Algorithm Identified Parameters

I am looking into Genetic Algorithms and it gave me the idea to use a genetic algorithm to train the parameters that are used by the Neural Network. This would help with the issue of what parameters you could use because the neural network would evolve into a set of fairly optimal parameters. Is this something that people are doing today? I am still fairly new to genetic algorithms, but this is a use-case that came to mind when I was learning about them.