The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) Assessing Data Quality at Shopify with Wendy Foster - #592
Sep 19, 2022
Wendy Foster, Director of Engineering & Data Science at Shopify, shares her insights on the shift from model-centric to data-centric AI. She discusses the critical role of data quality and freshness in enhancing merchant experiences. Wendy details challenges in product taxonomy and integrating images for identification, emphasizing collaboration. The conversation also touches on innovative inventory solutions and how Shopify's AI leverages user-provided labels for better customization. Get ready for a deep dive into the world of data-driven decision-making!
AI Snips
Chapters
Transcript
Episode notes
Early Career Focus on Model Optimization
- Wendy Foster recalls prioritizing micro-optimizations in model performance at a small startup.
- They focused less on data quality despite having a rich dataset, prioritizing engineering over impact.
Data Quality at Shopify
- Data quality, especially trustworthiness, reliability, and SLAs, is crucial at Shopify.
- Data governance ensures merchant data protection and emphasizes data coverage and freshness for modeling.
Data Innovation vs. Model Updates
- Shopify prioritizes constant data innovation over less frequent model changes.
- Model updates are driven by data changes, with occasional larger updates like incorporating image information.