System Design: Room Allocation and Ranking Service

Context and assumptions:

• The client sends a meeting request with: (a) time window or fixed start/end, (b) duration, (c) attendee count, (d) required equipment, and (e) preferred location/building.
• The service returns a top-N ranked list of candidate rooms (or room-time pairs when a window is provided) that can be immediately reserved.
• Hard constraints (capacity, equipment, availability) must be met; remaining factors are optimized via a score.

Requirements

1. API returns a ranked list of rooms (or room-time suggestions within the window) that satisfy hard constraints.
2. Propose a scoring function that integrates:
   • Historical room usage counts
   • Expected idle-time fragmentation
   • Attendee walking distance
   • Equipment match quality
   • Past no-show rates
   • Fairness across teams
3. Describe:
   • Data sources and freshness (online vs offline features)
   • Weight tuning strategy (rule-based vs learned model)
   • Safeguards against starvation (rooms/teams)
   • Online/offline evaluation
   • A/B testing plan