Abstract: Shopify’s Search and Discovery team is responsible for generating recommendations for millions of merchants that span many industries and countries. Specifically, when we think about content-based recommendations, we deal with product descriptions that vary in length, cleanliness and even coherence. In this talk, we will explore:
- The challenges we face when building content-based recommendation systems at Shopify.
- How we generated high-quality product embeddings using Universal Sentence Encoder (USE).
- Why we chose USE over other popular options such as BERT
- How we scaled our approach using Ray Actor Pools to generate recommendations for over 350M products.
- The impact of launching this new model to millions of merchants.
Very familiar with Python, Limited familiarity needed for Tensorflow and Ray Actorpools
Bio: Madhav is a Senior Data Scientist at Shopify where he focuses on building/evaluating recommendation systems. His role includes prototyping potential solutions and scaling them for production. Prior to Shopify, Madhav was a data science consultant where he focused on NLP projects for pharmaceutical companies. He then transitioned to Disney to develop personalized movie recommendations which sparked his passion for recommendation systems. In his free time, Madhav hosts free Q&A sessions for aspiring data scientists who are looking to get into this space.