Hotel Pricing System Design

Design question on hotel pricing (design an ML system/approach for pricing hotels). Asked in the design interview stage. The topic is intentionally open-ended; clarify the objective, users, marketplace constraints, and serving requirements before proposing an approach.

Address:

1. Define the pricing objective, prediction target, decision horizon, and required inputs.
2. Estimate scale, data volume, latency, freshness, and infrastructure needs using clearly stated assumptions.
3. Design the offline training and online serving architecture, including comparable-hotel retrieval, price prediction or ranking, and constraint-based re-ranking where useful.
4. Choose models and explain batch versus online inference.
5. Define offline and online evaluation, experimentation, monitoring, feature drift, and training-serving skew handling.
6. Discuss cold start, sparse booking outcomes, price changes, market shocks, and safe fallbacks.