Design photo and listing quality models
Company: Airbnb
Role: Machine Learning Engineer
Category: Machine Learning
Difficulty: medium
Interview Round: Onsite
Discuss how you would solve the following two machine learning product problems for a travel marketplace.
1. Improve booking performance by selecting the best cover photo to show for a listing or experience.
2. Identify high-quality listings in order to improve user experience through ranking, promotion, or quality enforcement.
For each problem, describe:
- problem framing and business objective
- training labels and their limitations
- useful features
- model choices
- offline and online evaluation metrics
- experimentation strategy
- cold-start handling
- bias, feedback-loop, and fairness risks
- deployment and monitoring considerations
Quick Answer: This question evaluates competency in machine learning product design and engineering, encompassing labeling strategy, feature engineering, model selection, evaluation metrics, deployment, monitoring, and fairness considerations for optimizing cover photo selection and identifying high-quality listings.