> Markdown version of [/jobs/ext/4631911-backend-engineer-search-all-genders](https://www.wearedevelopers.com/jobs/ext/4631911-backend-engineer-search-all-genders). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# Backend Engineer - Search (all genders)

- **Company:** About You
- **Location:** Hamburg, Germany
- **Experience:** Experienced
- **Contract:** Permanent contract
- **Skills:** Java (Programming Language), Artificial Intelligence, Airflow, BigQuery, Data Deduplication, Elasticsearch, Node.Js, Search Technologies, Data Streaming, TypeScript, Datadog, Large Language Models, HNSW (Hierarchical Navigable Small World), Backend, Kotlin, Low Latency, Search Engines, Google Gemini, Entity Resolution, OpenSearch
- **Published:** November 8, 2026
- **Apply:** https://startup.jobs/backend-engineer-search-all-genders-about-you-se-co-kg-10716829

## About the Role

* 3+ years of experience building scalable, low-latency backend APIs (Node.js/TypeScript preferred, or fast-converting from Java, Go, or Kotlin)
 * Real depth in Elasticsearch/OpenSearch (mappings, custom analyzers, percolators, scoring functions) beyond just basic query clients
 * Strong algorithmic intuition for ranked data (merging, deduplication, score blending, and latency trade-offs)
 * You're excited to bridge backend engineering and ML by serving model scores, embeddings, or LLM features in production
 * Fluent English communication skills and a pragmatic, data-curious mindset

Nice to Have

 * IR/NLP fundamentals (tokenization, multilingual search, relevance evaluation frameworks)
 * Vector search experience (k-NN/HNSW indices, SigLIP embeddings, recall vs. latency tuning)
 * LLM engineering (prompt caching, timeout budgets, Datadog LLM Observability)
 * Familiarity with BigQuery, dbt, or Dagster pipelines

## Description

* Develop the search engine behind ABOUT YOU, covering query understanding, entity resolution, and semantic vector search in Node.js/TypeScript
 * Blend OpenSearch relevance, embedding similarity, and behavioral re-ranking signals into unified search results
 * Build search-index pipelines (entities, synonyms, k-NN product vectors) streaming data from BigQuery into OpenSearch
 * Integrate LLMs and embedding models (Gemini, Vertex AI) under tight latency budgets with caching and fallback layers
 * Guard search relevance using golden lists, regression suites, and continuous A/B experiments
 * Shape architecture alongside our Tech Lead and collaborate closely with data analysts, engineers, and product managers