Design a User-Respecting Smart-Grid Controller
Company: OpenAI
Role: Software Engineer
Category: System Design
Difficulty: hard
Interview Round: Technical Screen
# Design a User-Respecting Smart-Grid Controller
Design a controller for an electrical grid connected to smart devices such as electric vehicles, refrigerators, air conditioners, and hospital equipment. When load is too high, the system should request or coordinate reductions while respecting that device owners retain final authority and devices often disconnect and reconnect.
### Constraints & Assumptions
- Safety-critical loads cannot be treated like deferrable household loads.
- A command may be delayed, rejected, or never acknowledged.
- The system must not infer consent merely because a device is online.
### Clarifying Questions to Ask
- What grid signal triggers intervention, and how quickly must load change?
- Which device capabilities and owner preferences are known?
- What guarantees apply to hospital or other critical equipment?
```hint Optimize under uncertainty
Treat each device response as uncertain and re-plan from measured load rather than assumed compliance.
```
### What a Strong Answer Covers
- Device registration, capabilities, priorities, preferences, and consent.
- Control-loop design, forecasting, dispatch, acknowledgments, and re-optimization.
- Offline/reconnect reconciliation and stale-command prevention.
- Safety, fairness, security, auditability, and fallback operation.
### Follow-up Questions
1. How would incentives be incorporated without disadvantaging users who cannot shift load?
2. What happens if connectivity returns after an overload event has ended?
Quick Answer: Design a resilient smart-grid controller that manages overload while honoring owner consent, critical loads, intermittent devices, and fairness.