Automated ML with PyTorch - any effort?

There ir quite expensive AutoML for Tensorflow, there is Auto-Keras for Keras/Tensorflow/etc. But what about automated ML for PyTorch? I have not heard about completed framework, but maybe there are some ongoing efforts to create such automated ML and maybe I can join? It would not be nice to start from scratch. And I don’t see future for PyTorch if not automated ML is supported.

Currently I am using PyTorch, because AllenNLP uses it and there are good libraries - neuron functions, gradiens, optimizers and so on. So - I hope that there are efforts in progress and that there will be future for PyTorch and it is wise to invest in PyTorch. Thanks.

And if there is no such effort, so - why there is no one? Maybe I can start such project but what obstacles and hardships should I take into account?