Airbnb·Software Engineer·Onsite - System Design / Architecture·Senior

Senior

Jun 2026

Summary

Airbnb system design round for a software engineer role, heavy ML focus. The whole session was basically one giant open-ended question about building a customer LTV system from scratch, which sounds manageable until you realize how many moving parts they actually want you to cover.

Questions Asked (1)

Design a complete machine learning system to estimate customer lifetime value on a platform. Cover everything from defining what LTV means for the business, to data pipelines, modeling choices, offline and online architecture, how scores get used downstream, and how you'd run experiments to validate impact.

System DesignProduct Analytics & MetricsA/B Testing & Experimentation

Author's notes

This is a beast of a question and I did not pace myself well.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying the business definition of LTV and how it will be used, then design the data pipeline and modeling approach, followed by offline/online architecture and downstream integration. Finally, outline an experimentation plan to validate the model's impact on business metrics.

Pro tip: Emphasize the importance of aligning LTV with business KPIs and ensuring the model is actionable; also highlight the need for continuous monitoring and retraining to adapt to changing customer behavior.

1. Define LTV and Business Objectives

Clarify what LTV means for the platform (e.g., predicted revenue over a horizon) and how it will be used (e.g., targeting, budgeting). Align with stakeholders on success metrics.

2. Data Pipeline and Feature Engineering

Identify data sources (transactions, engagement, demographics) and design ETL pipelines to compute features. Ensure data quality, freshness, and scalability.

3. Modeling Approach

Choose appropriate models (e.g., probabilistic, regression, deep learning) based on data characteristics and business needs. Validate offline with metrics like MAE, and consider interpretability.

4. Offline/Online Architecture and Integration

Design batch and real-time scoring systems, and integrate scores into downstream applications (e.g., CRM, marketing). Ensure low latency and scalability.

5. Experimentation and Validation

Plan A/B tests to measure impact of LTV-based interventions on business metrics. Define control/treatment, sample size, and success criteria.

Key Points to Mention

Business alignment: LTV definition should reflect company goals and be actionable.
Data quality and feature engineering: handle missing data, create time-based features, and ensure pipeline reliability.
Model selection and validation: consider trade-offs between accuracy and interpretability; use time-based cross-validation.
Scalable architecture: support both batch and real-time scoring with monitoring for drift.
Downstream integration: use LTV scores for personalization, targeting, and budget allocation.
Experimentation: design A/B tests with clear hypotheses, metrics, and guardrail metrics to measure incremental impact.

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.