MOC — Case Studies
Curated map of real-world search builds, grouped by the problem each team was solving. Field reports on how production search was built, scaled, migrated, and (sometimes) broken. For the flat list, see the Case Studies index.
Ranking & Learning to Rank in Production
- Airbnb - ML-Powered Experiences Ranking — ML ranking and the data flywheel
- Beyond Algorithms - Ranking at Scale at Booking.com — accommodation marketplace ranking; the work beyond algorithm choice (signals, leakage, sharded serving)
- Skyscanner - Learning to Rank for Flights — LTR in production over flight inventory
- Slack - Enterprise Message Search with LTR — enterprise/personal search; two-stage LTR, position-bias correction
- Kleinanzeigen - Vespa Migration for Homepage Feed — feed ranking after migrating to Vespa
Architecture & Platform Migration
-
Migration between Search Engines — topic hub: Solr/ES/OpenSearch/Vespa migration patterns (schema/analyzer parity, live-traffic sync, validation)
-
Amazon OpenSearch Service now offers AI-assisted migrations — AWS Migration Assistant; agentic Solr/ES/OpenSearch → OpenSearch
-
How I learned Vespa by thinking in Solr — cross-engine mental mapping (Solr → Vespa) as an onboarding aid
-
Netflix - Content Search Architecture — search architecture for a content catalog
-
Canva - Search Pipeline Modernization — modernizing a legacy search pipeline
-
Bonsai - Designing Search for a Relational Database — building search on top of a relational store
-
Kleinanzeigen - Vespa Migration for Homepage Feed — the migration itself, and why
-
Vinted - Migrating Search from Elasticsearch to Vespa — billion-item ES→Vespa migration; operational wins, Lucene analyzers in Vespa
Retrieval & Vector Search
- Reddit - Vector Database Selection — how to actually choose a vector database
- Uber Eats - Scaling Search for Food Delivery — geo-constrained retrieval and ranking at scale
Query Understanding & Quality
- Etsy - Search Quality and Query Understanding — query understanding in a long-tail marketplace
Cautionary Tales
- Zalando - Self-DoS via Facet Aggregation — when facet aggregation took search down