Which model should we use?
Answer it out loud before you open anything. The value of the flags below is in comparing them to what you actually said — including whether you asked about the data before naming a model.
A product manager for a marketplace asks the ML team to "build a model that predicts which listings will sell in the first week". The data is roughly 200,000 listings a month with title, description, photos, price, category, seller history and a sold-within-7-days label. The prediction feeds a seller-facing "improve your listing" hint and, later, maybe ranking. The PM asks in the kickoff: "so which model — XGBoost?"
What it is really testing
Whether the candidate can hold the question open. The wrong answer is any model name given before asking about the data, its size and modality, the latency budget, the interpretability requirement, the metric and the training and serving cost. The best candidates also ask what the prediction is *for*, because a seller-facing hint and a ranking signal want different things.