How do you handle disagreements with research scientists regarding the feasibility of deploying a specific model?
