Senior Data Scientist, Guest Travel Insurance (Algorithms)

Airbnb was born in 2007 when two hosts welcomed three guests to their San Francisco home, and has since grown to over 5 million hosts who have welcomed over 2 billion guest arrivals in almost every country across the globe. Every day, hosts offer unique stays and experiences that make it possible for guests to connect with communities in a more authentic way. The Community You Will Join: Airbnb is a mission-driven company dedicated to helping create a world where anyone can belong anywhere. Travel should feel safe—and AirCover is how we deliver on that promise. Through Guest Travel Insurance (GTI), we offer guests peace of mind at the moment of booking and throughout their trip. As a Data Scientist on AirCover, you’ll work at the intersection of insurance, personalization, and machine learning—building intelligent systems that help the right guest discover the right coverage at the right moment. You’ll join a tight-knit, high-output DS team that runs one of Airbnb’s most experiment-dense personalization roadmaps, partnering daily with product, engineering, operations, and legal to ship work that directly affects guest trust and revenue. The Difference You Will Make: We’re looking for a machine learning expert who is excited to own hard problems end-to-end—from prototype to production. You’ll have direct scope to contribute and lead across: - Package personalization & ML-based recommendation: Evolve rule-based guest segmentation into a full ML recommendation system that surfaces the right insurance (e.g., trip cancellation, accidental damage coverage, on-trip protection) to each guest based on purchase intent, trip attributes, listing signals, and user history. - Content personalization: Build models that rank and select benefit messaging for each guest—deciding which coverages to highlight, in what order, and with what framing—drawing on learnings from segmentation experiments and LLM-assisted content prototyping. - Intent modeling: Develop and productionize ML models (from gradient-boosted trees to deep learning) that predict a guest’s likelihood to value specific coverages, using structured booking data and unstructured signals. - Journey understanding and optimization: Leverage reinforcement learning to personalize across user journey, with understanding on user preferences on entry point, price, notification frequency, and trip characteristics - High-velocity experimentation: Design and run adaptive experiments to maximize learning within tight traffic constraints; sequence ERFs strategically to keep the personalization roadmap moving. A Typical Day: - Dig into experiment results to surface high-impact personalization opportunities; translate what you find into crisp scientific problem formulations that balance rigor with speed-to-learning. - Work closely with product managers, engineers, operations, legal, and privacy partners to align on ML requirements, de-risk design decisions, and gather requirements on explainability and compliance. - Hands-on develop, evaluate, and ship ML models and data pipelines at scale—batch and real-time, structured and unstructured—using Airbnb’s paved-path tooling and AI native mindset - Prototype and iterate quickly: turn a new idea into a working model in a prototype, get early signals from an experiment, then productionize what works. You move fast and don’t wait to be asked. - Present findings and proposals at team reviews and to technical, product, and executive stakeholders—making complex ML results legible without dumbing them down, and generating conviction on the roadmap ahead. - Stay current with the research community; draw on state-of-the-art advances in recommendation systems, LLMs, and personalization to raise the bar for what the team ships. Occasionally publish externally or present at conferences to advance Airbnb’s scientific standing. Your Expertise: - 5+ years of relevant industry experience (e.g., ML scientist, tech lead, junior faculty) and a Master’s degree or PhD with 2+ yrs in a relevant field. - Proven hands-on experience building and shipping personalization and recommendation systems at scale: strong intuition for feature engineering, user modeling, and the full ML lifecycle (training, serving, monitoring, iteration). Experience with LLMs, Computer Vision or content-understanding topics is a strong plus. - Strong fluency in Python and SQL; hands-on experience with TensorFlow or PyTorch, Airflow, and a data warehouse environment. - Deep understanding of ML algorithms (gradient-boosted trees, deep learning, optimization) and experiment design—including A/B testing, multi-armed bandits, and the practical constraints of running experiments at scale. Causal inference skills are a plus. - Exceptional communicator: you can make complex ML work legible to engineers, product managers, legal, and executives alike— written and verbal. You treat communication as a core part of the job, not an afterthought. - Self-directed and pas

5+ years of relevant industry experience (e.g., ML scientist, tech lead, junior faculty) and a Master’s degree or PhD with 2+ yrs in a relevant field. Proven hands-on experience building and shipping personalization and recommendation systems at scale: strong intuition for feature engineering, user modeling, and the full ML lifecycle (training, serving, monitoring, iteration). Experience with LLMs, Computer Vision or content-understanding topics is a strong plus. Strong fluency in Python and SQL; hands-on experience with TensorFlow or PyTorch, Airflow, and a data warehouse environment. Deep understanding of ML algorithms (gradient-boosted trees, deep learning, optimization) and experiment design—including A/B testing, multi-armed bandits, and the practical constraints of running experiments at scale. Causal inference skills are a plus. Exceptional communicator: you can make complex ML work legible to engineers, product managers, legal, and executives alike—written and verbal. You treat communication as a core part of the job, not an afterthought. Self-directed and passionate about solving hard problems.

Own hard problems end-to-end—from prototype to production. Contribute and lead across: package personalization & ML-based recommendation to surface the right insurance to each guest; content personalization to rank and highlight coverages; intent modeling to predict likelihood guests value specific coverages; journey understanding and optimization using reinforcement learning to personalize across user journeys; high-velocity experimentation to maximize learning and carefully sequence experiments. Typical Day: analyze experiment results to surface impactful personalization opportunities and translate findings into scientific problem formulations; collaborate with product, engineering, operations, legal, and privacy to align ML requirements and ensure explainability and compliance; hands-on development, evaluation, and deployment of ML models and data pipelines at scale (batch and real-time); prototype quickly, iterate based on early signals, then productionize; present findings to team, product, and executive stakeholders; stay current with research and potentially publish or present externally.