What problem does it solve? Preparing for senior data-engineering data-modeling rounds is hard to practice alone because the real signal is how a candidate handles an under-specified prompt that changes mid-design. This Skill runs a live role-play interview where an interviewer persona hands over an open-ended analytics-backend design prompt, withholds key facts until asked, and injects escalating complications, then delivers a rubric-scored coaching debrief. ## Core Features & Use Cases - Live interviewer role-play: Two calibrated personas (Staff DE and Principal DE/Architect) run a ~50-60 minute round covering requirements gathering, conceptual modeling, datastore selection, physical design, and an escalation ladder. - Escalation playbook: Seven reusable complications (real-time SLA, exact billing, schema evolution, GDPR deletes, 10x skew, multi-tenant isolation, cost pressure) injected one at a time to test whether the model adapts or breaks. - Flagship OpenAI analytics case plus question bank: Ships with an OpenAI product-data scenario and six alternate domains (ride-sharing, ad-tech, IoT, e-commerce, fintech ledger, SaaS metering). - Scored coach debrief: Six-dimension senior-bar rubric (requirements, grain, datastore, physical design, quantitative reasoning, escalation resilience) with per-phase micro-feedback and an optional no-feedback exam mode. - Use Case: A senior data engineer asks to "mock the OpenAI data model question", practices stating fact-table grain and defending a lakehouse-plus-serving-store architecture, then receives a debrief naming the single highest-leverage fix before their real onsite. ## Quick Start Ask Claude to start a mock data model interview using the OpenAI analytics case with per-phase feedback enabled.
