[flooo]

I’m a researcher working in the fraud detection domain.

Do you have some pointers on scaling GNNs to such large problems?