MOC — Architecture and Search Team
Map of content covering search system architecture and the human side of search: team structure, process, and organizational practices.
System Architecture
Concepts
- Concept: Search Architecture — end-to-end pipeline: ingestion → retrieval → ranking → serving
- Concept: Retrieval Pipeline — multi-stage retrieve → rerank
- Concept: Knowledge Graph Search — entity graph + vector integration
Case Studies
-
Canva Search Pipeline Part I — motivation: big ball of mud, 20k rps
-
Canva Search Pipeline Part II — solution: isolated, stateless, immutable phases
-
Optimizing Search at Uber Eats — geosharding (H3), layout optimization, ETD indexing; Janani Narayanan & Karthik Ramasamy
-
The Day Our Own Queries DoSed Us - Zalando Search — Maryna Kryvko; layered architecture; self-DoS via facet aggregation
Search Team Organization
Process
Process
- Query Triage - The Secret Weapon for Search Relevance — Rubinstein: cross-functional triage meetings
- The Launch Review — go/no-go decision process for algorithm changes
- Search Relevance for Understaffed Teams — Turnbull: focus on NOW, not future headcount
- Getting Started on Search Relevance for the Understaffed Search Team — comprehensive strategy guide for lean teams
- Understaffed Search Team — topic: high-ROI techniques for resource-constrained teams
Key Figures in Search Teams
- Daniel Tunkelang — search leadership, strategy, philosophy
- Doug Turnbull — search engineering, LTR, understaffed teams
- James Rubinstein — search product management, measurement
- Andreas Wagner — searchhub.io, Three Pillars of Search Quality
Concepts
Evaluation and Quality Process
Tools and Methods
- Concept: Search Evaluation — offline vs. online; judgment-based vs. behavioral
- Concept: Judgment Lists — (query, document, grade) triples
- Concept: A-B Testing for Search — online experimentation
- A-B Testing for Search is Different — why session-level analysis matters
Search Team Roles
- Search Relevance Engineer: LTR, ranking signals, evaluation harness
- Search Product Manager: query triage, launch reviews, business alignment
- Search Data Scientist: metric design, experiment analysis, judgment collection
- Search UX Researcher: user intent research, usability testing
Related MOCs
- MOC - Search Quality Assurance and Query Understanding
- MOC - Ranking and Retrieval
- MOC - Case Studies
Additional Articles — Case Studies & Architecture
- Netflix Elasticsearch Indexing Strategy in AMP — alias-based index management, mapping design, refresh tuning
- Twitter Search Stability and Scalability — Earlybird architecture, circuit breakers, real-time indexing at scale
- Thoughts about Managing Search Teams — combine product+engineering, data-driven, incremental execution
- Search Product Management - The Most Misunderstood Role — search PM must understand indexing, ranking, and experimentation