THE COMPANY INDEX TRACKED BLOG
Ideas, decisions, and lessons from the team.
shopify.engineering (opens on the source site)LinkedIn XShopGym creates realistic online stores and generates shopping tasks based on each store’s products and features. Developers can then repeatedly test shopping agents under consistent conditions.
Read at the sourceWe built Helix to help LLMs rebuild the Shopify app in Swift and Kotlin, using small checkpoints and strict quality gates to keep the code shippable.
Read at the sourceWe migrated the Shop app from React Native to Swift and Kotlin. Assisted by AI, the team was able to go from a proof of concept to a fully rebuilt native app published in the app stores in just 12 weeks.
Read at the sourceCoding agents changed what it costs to build mobile apps twice. Here’s why Shopify is moving from React Native back to Swift and Kotlin.
Read at the sourceGisting compresses context into a set of learned tokens, preserving its quality while making the model faster and cheaper.
Read at the sourceWe rebuilt our mobile end-to-end testing framework with a strict API and computer vision, raising test stability drastically.
Read at the sourceHow we compress production failures into model weights every day, beat frontier-model quality, and cut serving costs 96%.
Read at the sourceWe built an agentic code review and test oracle harness that discovers vulnerabilities, proves them with real tests, and provides Shopify-tuned fixes.
Read at the sourceWe moved five high-traffic checkout extensions to remote-dom and Polaris web components, cutting bundle sizes drastically and making checkout faster.
Read at the sourceHow a Hack Days team built a music player, custom GLSL visualizers, and an artist toolkit for storefronts, all in three days.
Read at the sourceBuilding an LLM-powered pipeline to match product listings across millions of merchants into a unified catalog that AI agents can search.
Read at the sourceProduction training data only captures successful queries; it can't teach a model when to say no. We built an automated curation pipeline using LLM judge consensus to close that gap.
Read at the sourceWhat it took to ship our Slack-native agent River, lessons learned, and the substrate that runs beneath it. Co-authored by River.
Read at the sourceWe fine-tuned Qwen3-32B into a tool-calling agent that generates Flow automations from natural language—faster, cheaper, and more accurate than the frontier model it replaced, with a weekly retraining flywheel built on real merchant data.
Read at the sourceThe technical blueprint for an AI-powered mirror that sees customers, analyzes their appearance, and delivers personalized product recommendations in real-time.
Read at the sourceConventional GraphQL execution uses depth-first traversal that incurs many hidden costs. We questioned why, rewrote it in a faster breadth-first manner, and saw dramatic results.
Read at the sourceGPUs are annoying. Shopify uses SkyPilot to make them less so: one YAML file, multiple clouds, clean development ergonomics.
Read at the sourceTangle saves months of compute time, makes every experiment automatically reproducible, and allows teammates to share computation without coordination.
Read at the sourceFrom March to October we simulated traffic tsunamis, injected chaos, and fixed every bottleneck before our merchants needed us most.
Read at the sourceLast year, over 875 million people bought items from Shopify merchants. Building on our prior Vision Language Model-based product classification, this post explores how AI agents are evolving the taxonomy itself.
Read at the source