Cooking up a new search system: Recipe search at Cookpad
summary
Cookpad replaces a brittle legacy search with a new event-driven, Python-centric stack (Kubernetes, Kafka, Python, Elasticsearch, ML) over two years, keeping live traffic via a strangler fig migration with shadowing, progressive rollouts and auto-retry to the old system. Responsibility shifts to two cross-functional search teams plus a platform team; experiments run weekly, targeting segments like single-ingredient queries, and deliverables include a Django/Postgres admin UI, search filters, faster responses, measurable uplifts in cook-through rate, and only a few incidents per year.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 4:51The Search Problem
- 12:40Query Segments
- 16:48Qualitative vs Quantitative
- 18:20Autonomy Empowerment
- 21:38Team Topologies
- 25:47Search Engineering
related talks
44mBuilding a Modern Search Engine: The Road Ahead
33mDeveloping for Multi-Component Environments While Keeping Your Sanity
44mCooking up Better Front-end Applications with Inspiration From the Culinary World
31mNext-Gen Search: How Uber and the OpenSearch Community Built a Cloud Native 3.0
32mGrowing Pains, Portal Gains: The Evolution of Skyscanner
19mPivoting to React, at scale
51mCode spelunking in the All New Basecamp
45mMigrating Pinterest profiles to React