Etsy uses deep learning to rank search results and tailor the shopping experience to each buyer across its marketplace of 100+ million listings. The system learns from signals like clicks, purchases, and browsing history to decide which items each person sees first.

Etsy migrated from older tree-based ranking models to a deep learning neural network, reducing model training time by 8x using GPUs and TensorFlow on Google Cloud. The system scores listings based on relevance, listing quality (clicks, favorites, purchases), seller service quality, and personalization signals derived from each buyer's history. In 2025, Etsy further upgraded search with large language models that score items across broad interest ranges and power LLM-generated buyer profiles, enabling personalized query rewriting for more relevant results. Etsy's FY2024 annual report (10-K) formally discloses the use of AI, generative AI, and machine learning across search and discovery surfaces.