Design a deployment strategy for a model that needs to run inference on an edge device with limited compute power for industrial operations.
