A team is starting a supervised ML project with a measurable business outcome. Which high-level sequence best reflects a sound development lifecycle?

1. Deploy first, use production outcomes as training and test data simultaneously, and define the business objective last.
2. Collect data, publish every model revision automatically, and stop monitoring once the endpoint responds.
3. Define the objective, prepare suitable data, train a candidate, evaluate on held-out data, deploy after validation, and monitor outcomes.
4. Choose the largest model, skip data review, and approve deployment solely from training accuracy.