Databricks product spec: pick a capability and define scope, requirements and metrics

Choose one Databricks product and write a product requirements document (PRD) for it. Be ready to walk the interviewer through each section and defend your choices. ```hint Pick a problem you can defend Choose a capability where you can name the specific Databricks user, the pain they hit today and the evidence you would use, before you write a single requirement. ``` ### Clarifying Questions - Should the PRD describe a new capability inside an existing product, such as Databricks SQL, Unity Catalog or MLflow, or an entirely new product? - Who reads the document first: the engineering team that will build it, or the leaders deciding whether to staff it? - Is this a written exercise with a length limit, or a live walk-through of the outline? - Since I have no internal data, may I state the supporting evidence (usage data, customer interviews) as hypotheses to validate? ### What a Strong Answer Covers - A reasoned choice of product and problem, tied to a specific platform user such as a data engineer, analyst, ML engineer or workspace admin. - A problem statement backed by evidence or testable hypotheses, with explicit goals and non-goals. - Prioritized requirements written as user-visible behavior, with the core user flow and the dependencies on other parts of the platform. - Success metrics and guardrails that suit a usage-billed data platform. - Risks, a staged rollout plan and the open questions the team still has to answer. ### Follow-up Questions - If your feature would lower some customers' compute consumption, and with it near-term revenue, how would you argue for it to leadership? - What would you cut to get a first version to design partners quickly, and what would you refuse to cut? - How would your requirements change for customers in tightly regulated industries?