How to compose TensorBoard into pipeline during training

I have tried MLBlocks, MLPrimitives, Orion to train neural network models, but it seems like there is no way to add some custom operations, such as TensorBoard.

BTW, I'm curious about roadmap of MLBlocks. Once I tried this, it didn't save my time actually, due to lots of annotation work. But I was wondering it's good for AutoML. TF/Pytorch do have good AutoML tools, so I'm not sure for now.