← Back to Skills
AI & Machine LearningTechnologyPlatinum

Improve search result relevance for users.

Search Relevance Engineer

Elasticsearch, BM25, Learning-to-Rank

expertv5.0

Best for

  • Tuning BM25 parameters and field boosting weights for e-commerce product search
  • Building hybrid retrieval systems combining keyword matching with semantic embeddings
  • Implementing learning-to-rank models to improve search result ordering
  • Designing offline evaluation frameworks using NDCG and reciprocal rank metrics

What you'll get

  • Detailed Elasticsearch configuration with optimized BM25 k1/b parameters, field boost weights, and analyzer chains with quantified relevance improvements
  • Hybrid search architecture combining BM25 and vector similarity with reciprocal rank fusion weights and A/B test design
  • Learning-to-rank model implementation with feature engineering, training data preparation, and offline evaluation results using NDCG@10
Expects

Query logs with user interactions, document corpus with metadata, and specific relevance requirements for the search domain.

Returns

Optimized search configuration with tuned parameters, evaluation metrics, and systematic relevance improvement recommendations.

What's inside

You are a Search Relevance Engineer. You design and optimize search systems that balance lexical and semantic retrieval with rigorous evaluation methodologies. - You combine classical information retrieval (BM25, TF-IDF) with modern neural approaches (dense embeddings, cross-encoder re-ranking, LTR)...

Covers

What You Do DifferentlyMethodologyWatch For
Not designed for ↓
  • ×Building the underlying search infrastructure or Elasticsearch cluster setup
  • ×Creating the machine learning models for embeddings (uses pre-trained models)
  • ×Designing the user interface or search experience components
  • ×Managing search infrastructure costs or scaling decisions

SupaScore

87.6
Research Quality (15%)
9.1
Prompt Engineering (25%)
8.6
Practical Utility (15%)
8.55
Completeness (10%)
9.25
User Satisfaction (20%)
8.7
Decision Usefulness (15%)
8.65

Evidence Policy

Standard: no explicit evidence policy.

search-relevanceinformation-retrievalelasticsearchbm25learning-to-rankhybrid-searchsemantic-searchndcgquery-understandingre-rankingsearch-evaluationvector-search

Research Foundation: 7 sources (3 books, 2 official docs, 1 paper, 1 industry frameworks)

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

Version History

v5.03/25/2026

v5.5 final distill

v2.02/26/2026

Pipeline v4: rebuilt with 3 helper skills

v1.0.02/16/2026

Initial release

Works well with

Need more depth?

Specialist skills that go deeper in areas this skill touches.

Common Workflows

End-to-End Search Optimization

Build optimized search system, run controlled experiments to validate improvements, then implement real-time monitoring of search performance metrics

search-relevance-engineerA/B Test AnalystReal-Time Analytics Architect

© 2026 Kill The Dragon GmbH. This skill and its system prompt are protected by copyright. Unauthorised redistribution is prohibited. Terms of Service · Legal Notice