Machine Learning Engineer interview
at Airbnb.

Tell-me-about-a-time stories on ownership, conflict, and impact — scored against this company's real values.

1. 01 How does Airbnb's mission of belonging shape the way you'd approach fairness and bias in a ranking or pricing model you build?
2. 02 Tell us about a time you were a host to a partner team by making your model's behavior transparent and explainable to non-ML stakeholders.
3. 03 Describe a time you embraced the adventure and shipped an ambitious ML system despite incomplete data or uncertain payoff.
4. 04 Walk us through a time you scrappily acted like a Cereal Entrepreneur to ship an ML feature with minimal infrastructure, and how you iterated.

The hard screen for the craft itself — talked through out loud, not whiteboarded in your head.

1. 01 Given event-level booking and search logs, write code to engineer features for a model that predicts guest booking likelihood.
2. 02 A deployed search-ranking model suddenly performs worse in production than offline; how would you debug the discrepancy?
3. 03 What loss functions and evaluation metrics would you choose for a learning-to-rank model for Airbnb listing search, and why?
4. 04 How would you handle the cold-start problem for newly listed homes that have no booking or review history in a ranking model?

Open-ended design of the systems this company actually runs, with the interviewer probing your tradeoffs.

1. 01 Design Airbnb's machine-learned search and ranking system for listings, covering features, candidate retrieval, ranking, and online serving latency.
2. 02 Design a dynamic pricing-suggestion model for hosts, including data sources, training, guardrails, and monitoring.
3. 03 Design a trust-and-safety fraud-detection system that scores reservations in real time without blocking legitimate guests.
4. 04 Design a host-guest matching and personalized recommendation system, addressing two-sided marketplace tradeoffs and feedback loops.