Linked Presentation: BLAS-on-flash: An Efficient Alternative for Large Scale ML Training and Inference?
