Implement NASNet (Neural Architecture Search) in TensorFlow?


I am quite happy with some results I got, but I am still wondering, if finding the right Algorithm and Parameters etc. could be made easier. Then I came a cross a concept which is pretty old I guess from Google, where they developed a concept in which the machine searches and optimizes for the best way to set up a neural network itself, called NASNet. Is this possible with TensorFlow? If yes, also in Node Js or only in Python and how?

Thank you a lot!