Search Engineer
Search Engineers build and optimize the search infrastructure that powers product search, enterprise search, and information retrieval systems. Their daily English covers discussing relevance tuning strategies, presenting A/B test results for ranking changes, writing architecture documents for index design, and explaining recall/precision trade-offs to product teams. This path covers the precise technical vocabulary of information retrieval and search systems.
Topics covered
- Information retrieval theory
- Relevance & ranking
- Semantic/vector search
- Query understanding
- Search quality evaluation
- Scaling search infrastructure
Vocabulary spotlight
4 terms every Search Engineer should know in English:
A data structure that maps terms to the documents containing them — the core of most full-text search engines, enabling fast lookups of which documents contain a given word
"The inverted index allows us to find all documents containing a search term in milliseconds, even across billions of records."
Best Match 25 — a probabilistic ranking algorithm that scores documents by term frequency and document length normalisation, widely used as a strong lexical search baseline
"Our initial ranking used BM25, which performed well on keyword queries but struggled with synonyms."
The proportion of relevant documents that a search system returns — a high-recall system returns most relevant results but may include many irrelevant ones
"We improved recall by adding synonym expansion, but precision dropped as more irrelevant results appeared."
A dense vector representation of text (or other data) in a high-dimensional space, where semantically similar content is positioned close together — the foundation of semantic/vector search
"We switched to a bi-encoder embedding model to enable semantic search that can match user intent beyond exact keywords."
📚 Vocabulary Reference
Key terms organised by category for Search Engineers:
Core IR
Relevance & Ranking
Semantic Search
Evaluation
Recommended exercises
Real-world scenarios you'll practise
- Writing a search relevance tuning document: explaining how you used offline evaluation metrics (NDCG, MRR) to guide ranking changes
- Presenting an A/B test of a new ranking model to the product team: explaining precision/recall trade-offs in non-technical terms
- Architecting hybrid search: explaining to stakeholders why you are combining BM25 and vector search instead of choosing one
- Documenting the index schema design decision: why you chose certain field mappings, analyzers, and shard routing strategy