TechPrufer
Get started
Shopify
Shopify logoShopify · Feature trail

Recommender Systems for Merchants

6 posts · Jan 2019 – Feb 2026

What this trail covers

Shopify has developed and enhanced recommender systems to provide personalized insights, actions, tools, and resources to merchants. Initial efforts focused on leveraging collaborative filtering and user-item interactions to predict item relevance. This post details the development of a foundational generative recommender system that learns directly from raw event sequences, operates within real production latency constraints, and delivers measurable impact at scale. Key advancements include architecture decisions, training strategies, and optimized CUDA kernels for efficient training and serving. The system uses time-aware attention with rotary encoding and relative attention bias, improved negative sampling strategies (shared negatives, positive-aware hard negatives), and a boosting-inspired approach for incremental recall. Online A/B testing showed positive movement in key metrics like shop orders, click-through rate, and conversion rate. Future work explores moving from product ID space to token space through semantic IDs for reduced embedding table dependence and better integration of text queries and assistant interactions.

All Shopify trails
  1. The generative recommender behind Shopify's commerce engine (2026) - ShopifyThis post details the development of a foundational generative recommender system at Shopify. Key technical contributions include: training an autoregressive model with a causal mask on raw event sequences; incorporating time as a first-class signal via RoPE-inspired rotary encoding and relative attention bias; implementing advanced negative sampling strategies (shared negatives, positive-aware hard negatives) to improve retrieval quality and system efficiency; and using a boosting-inspired approach to achieve incremental recall for the ensemble. Optimized CUDA kernels were developed for efficient training and serving, leading to significant speedups. Online A/B testing demonstrated improvements in shop orders, click-through rate, and conversion rate.
  2. Augmented commerce: Machine learning at Shopify (2025) - ShopifyThis post details the application of machine learning at Shopify, expanding on recommender systems for merchants. It introduces the use of finetuned Qwen multimodal models for product classification and enrichment, fraud detection models inspired by Feature Space, LLaMa and general models for the Sidekick merchant assistant, Nomic embeddings for product catalog vector representations, tabular transformer models for GMV forecasting, HSTU architecture for behavior modeling, and small language models for query rewriting. It also highlights investments in GPU infrastructure (GCP, Nebius), partnerships for GPU computation acceleration (CentML), data annotation services (Toloka), and early access to frontier models.
  3. Shopify's Playbook for Scaling Machine Learning - ShopifyThis post details Shopify's playbook for scaling machine learning, focusing on a pragmatic, step-by-step guide applicable across any domain. It outlines five key steps for starting from zero: identifying a problem worth solving (e.g., order fraud detection), ensuring sufficient and accessible data, identifying downstream dependencies, understanding existing solutions to set benchmarks, and optimizing for product outcomes over pure model scores. It then details the 'Zero to One' phase, emphasizing the importance of well-defined pipelines for training and testing, and strategic model deployment decisions based on volume and user commitment. The 'One to One Hundred' phase focuses on building trust in models through input/output reconciliation, production backtesting, and continuous monitoring, and on scaling model building workflows by investing in data engineering practices and efficient deployment strategies. The post also touches upon the application of AI for complex routing problems and the development of AI assistants.
  4. 5 Steps for Building Machine Learning Models for Business - ShopifyThis post outlines a five-step business-centric approach to building and productionizing machine learning models. It emphasizes asking if ML is the right solution, starting with simple models and heuristics, measuring impact before optimizing, and having a plan for iteration. Specific techniques discussed include using simple models like linear regression, composite indices, and keyword matching as baselines, and the importance of tying performance metrics to business impact. It also highlights the need to understand model tradeoffs through confusion matrices and residual distributions.
  5. The Evolution of Kit: Automating Marketing Using Machine Learning - ShopifyThis post details the evolution of Kit from a rule-based system to an AI-driven assistant. It describes the transition from heuristic rules for marketing campaign configuration to machine learning models for predicting optimal budget (regression) and likelihood of sales (classification). The post outlines the architecture for training and prediction flows, including feature engineering using Apache Spark, model training on Google Cloud's ML Engine, and real-time prediction using Apache Kafka. It also introduces a new use case for acquiring visitors to build lookalike audiences, requiring a shift to real-time prediction.
  7. How Shopify Uses Recommender Systems to Empower Entrepreneurs - ShopifyThis post details the design and implementation of Shopify's recommender system platform, focusing on collaborative filtering (CF) using user-item interactions. It explains how user preferences are represented as vectors and treated as a binary classification problem to predict item relevance. The post highlights the use of linear models like LRec for learning item-item similarity matrices, enabling task-parallel training for scalability and interpretability through model coefficients. It also discusses the importance of performance metrics beyond precision, such as diversity and serendipity, and how user features can enrich the model to address cold-start problems.