This is a dedicated watch page for a single video.
A retail company has developed a recommendation ML model to enhance customer experience on its platform. Before fully deploying the model in production, the company wants to perform online validation by routing 20% of the traffic to the new model while the remaining traffic continues to use the existing model. Which solution will implement the online validation with minimal operational overhead?