Staff Data Scientist - Inference, Trust
San Francisco, United States
Airbnb is a mission-driven company dedicated to helping create a world where anyone can belong anywhere. It takes a unified team committed to our core values to achieve this goal. Airbnb's various functions embody the company's innovative spirit and our fast-moving team is committed to leading as a 21st century company.
About the position:
As part of the Anti-Discrimination team, you will be a part of a groundbreaking group of technologists designing and building scalable and robust systems to detect and support the mitigation of inequities across the entire platform. Those inequities are discrepancies in the product experience that may be correlated with race or other demographic attributes, and are the result of explicit or implicit discrimination on an individual level (e.g., between hosts and guests), as well as broader, systemic inequities.
You will be deeply involved in the technical details of expanding on the work we started with Project Lighthouse, in close collaboration with product, engineering and operations teams. Your work will help ensure that every product team at the company is empowered with the insights necessary to address inequities within their features and surface areas, head-on -- ultimately enabling us to build a product that fosters true belonging amongst members of the Airbnb community.
Airbnb's inference data scientists understand the complications of running A/B tests in a two-sided marketplace. They often conduct power analyses, clearly articulate stopping criteria, and use statistical models to improve inference. They also lead in developing analyses strategies for situations when A/B tests are infeasible. How do we understand the impact of something new? In this role, you will be deeply involved in defining, measuring and mitigating potential gaps in the Airbnb experience. Your expertise in causal inference as well as passion for the subject matter will help you succeed in tackling this unstructured and important problem.
The work of making our platform one that is more just and equitable is challenging, but critical. We’re looking for talented data scientists who are ready and excited to take on this important and unsolved problem.
- Designing and implementing analyses that measure and help mitigate potential inequities within the Airbnb user experience.
- Defining and evaluating key metrics in an unstructured problem space. Developing and communicating a robust understanding of what moves these metrics and why.
- Developing and implementing research and auditing methodologies to understand the impact of machine learning with regards to fairness.
- Working closely with a range of stakeholders, including external civil rights and privacy organizations and academics, to improve our technology.
- Investigating challenging questions around the user experience to understand the voice of our users at scale.
- Partnering with Engineering to improve data quality.
- Empowering the company to answer questions by building high-quality ground truth data products.
- Contributing to the fairness in the technology research community through publishing findings.
- Executing projects with great complexity and a longer time horizon than is typical on most product teams. Our projects can take 6-24 months to complete, and, as a result, require a sustained level of commitment and attention, as well as thorough, thoughtful, and rigorous documentation.
- 8+ years of relevant experience (6+ years with advanced degree)
- Expert in at least one programming language for data analysis (Python or R)
- Experience with experimentation and machine learning techniques
- Experience in schema design and high-dimensional data modeling
- Proven ability to succeed in both collaborative and independent work environments
- Passion for high data quality and scaling data science work
- A recognition of the problem we’re trying to solve, and curiosity about it. You recognize that issues of bias, discrimination, and systemic inequities are real, measurable problems, and you are eager to learn more about them.
- Strong skills in SQL and distributed system optimization (e.g. Hadoop, Hive, Spark, Presto)
- Experience with an ETL framework like Airflow
- Experience with causal inference and structural models
- Experience with academic research in fairness in technology
- Experience with technical mentorship of other data scientists
- Experience with prototyping and writing production code