[Insert Job Description Here]

Conversion data is the foundation for ads reporting and ads ML modeling at Pinterest. Making sure our conversion data carry a high signal-to-noise ratio of users’ commercial intent is critical for Pinterest to deliver business value to advertisers. We are seeking an experienced data scientist to help us shape the strategy of conversion data quality and develop solutions for enhancing the data foundation for Pinterest’s ads system. In this role, you will work on a variety of exciting data science challenges in this new frontier, such as how to define and evaluate data quality, how to establish causal signals from observational data, how to unlock the potential of ML model performance through the lens of data-centric AI, etc.

What you’ll do:

• Conduct deep-dive analysis to explore the patterns in real-world user conversion data, identify a variety of data quality issues and quantify their impact on our ads system.
• Analyze causal connections between users’ online behaviors and their conversions, uncover causal signals in data, and develop metrics to measure the effectiveness of ads.
• Design and prototype automatic solutions that detect quality issues in our conversion data, develop high-signal training labels and features to boost ads ML model performance, and collaborate with engineers to integrate them into our ads system.
• Collaborate closely with Product and Engineering teams as a thought partner, provide data-informed recommendations that shape the future roadmap and drive the long-term strategy for building a high–quality conversion data foundation. 

What we’re looking for:

• 5+ years of hands-on experience as a data scientist, applied scientist or machine learning engineer.
• Proven ability to apply scientific methods to solve real-world problems on web-scale data.
• Proficiency in SQL/Hive. Expertise in at least one scripting language (ideally Python/R).
• Strong business and product sense: delight in shaping vague questions into well-defined analyses and success metrics that drive business decisions.
• Excellent communication skills: able to communicate findings with leadership and product teams. 
• Experience in causal inference and measurement. Familiarity of measuring causal effect by both online experimentation and observational study. 
• Experience in data quality related work, including quality evaluation, anomaly detection, data cleaning, data governance, etc. 
• Proven ability to design and implement machine learning models, ideally with biased/incomplete/noisy training data. Deep understanding of feature design and model evaluation is strongly preferred.
• Proven ability to independently explore an ambiguous domain and shape the product roadmap and strategy by analytical efforts.

Relocation Statement:

•  This position is not eligible for relocation assistance. Visit our PinFlex page to learn more about our working model.

#LI-NM4

#LI-HYBRID

#LI-HYBRID or #LI-REMOTE