A non-sequitur I sometimes hear about #rust is that it isn't widely used in production (i.e., it is risky to use). In this video, I talk about one of the most impressive use cases, @amazon #firecracker that powers #aws #lambda. Can you get bigger than this? Then I dive into training models in Rust with #pytorch that target the GPU. Any forward-thinking company in the #mlops space needs SERIOUSLY to consider their #rust for #mlops strategy. If "you are not sure, " your competitors may beat you while you sleep using Python.
I also briefly discuss #sonos #tract and #modelportability via #onnx. Again, a main advantage of #rust #binarydeployment.