Hybrid Search Architect

Design production-grade hybrid search systems that combine BM25 keyword search with vector-based semantic retrieval, multi-stage reranking pipelines, and RAG retrieval layers for optimal relevance and recall.

Gold

v1.0.00 activationsAI & Machine LearningTechnologyexpert

SupaScore

Research Quality (15%)

8.5

Prompt Engineering (25%)

8.5

Practical Utility (15%)

8.5

Completeness (10%)

8.5

User Satisfaction (20%)

Decision Usefulness (15%)

8.5

Best for

▸Designing e-commerce search that combines product name matching with semantic similarity
▸Building RAG systems that balance exact citation retrieval with conceptual document matching
▸Optimizing knowledge base search for customer support with both keyword and intent understanding
▸Creating code search systems that find exact function names and semantically related implementations
▸Architecting multi-stage retrieval pipelines with BM25 candidate generation and neural reranking

What you'll get

●Multi-stage architecture diagram with BM25 candidate retrieval, dense vector scoring, RRF fusion weights, and cross-encoder reranking with specific latency budgets
●Embedding model comparison matrix with MTEB scores, dimensionality trade-offs, and domain-specific fine-tuning recommendations
●Production deployment guide with vector database selection, indexing strategies, and A/B testing framework for relevance optimization

Not designed for ↓

×Simple keyword-only search implementations without semantic components
×Training embedding models from scratch or fine-tuning transformers
×General database query optimization unrelated to search relevance
×Building recommendation systems based on collaborative filtering

Expects

Detailed search requirements including corpus type, query patterns, latency constraints, and relevance priorities with sample queries and expected results.

Returns

Complete hybrid search architecture with specific technology stack recommendations, fusion strategies, reranking pipelines, and implementation guidance with performance benchmarks.

Evidence Policy

Enabled: this skill cites sources and distinguishes evidence from opinion.

hybrid-searchvector-searchbm25semantic-searchrerankingcross-encoderreciprocal-rank-fusionrag-retrievalembedding-modelsinformation-retrievalrelevance-tuningquery-understandingdense-retrievalsearch-architecture

Research Foundation: 10 sources (2 academic, 6 paper, 2 official docs)

This skill was developed through independent research and synthesis. SupaSkills is not affiliated with or endorsed by any cited author or organisation.

Version History

v1.0.02/15/2026

Initial release

Works well with

Embedding Space ArchitectPlatinum LLM Evaluation Framework DesignerGold RAG Architecture DesignerGold Search Relevance EngineerPlatinum Vector Database OptimizationGold

Need more depth?

Specialist skills that go deeper in areas this skill touches.

Elasticsearch Query Optimization ExpertGold Embedding Space ArchitectPlatinum LLM Fine-Tuning StrategistGold

Common Workflows

Production RAG System Development

Design hybrid retrieval, integrate with RAG pipeline, then build evaluation framework for end-to-end relevance testing

hybrid-search-architect→RAG Architecture Designer→LLM Evaluation Framework Designer

Activate this skill in Claude Code

Start Free to Activate This Skill