Senior RAG Engineer

Newcode; United States; Remote; USD 150,000 - 190,000; Full time

Benefits offered by this job

Fully remote work in eligible regions

Job summary

Newcode is seeking a retrieval systems architect to own the end-to-end pipeline for extracting evidence from large confidential documents. You will design document processing, chunking, embedding, and retrieval workflows that feed language models in professional settings.

You will combine semantic search with exact-term retrieval, implement ranking, and set quality metrics such as citations accuracy. This remote, full-time role emphasizes independent work and impact across the product.

Qualifications

• Strong experience building retrieval or search systems for large document collections.
• Experience evaluating search relevance with BM25, NDCG, or recall@k.
• Ability to build search abstractions across multiple backends.
• Comfort working independently as priorities evolve.
• Authorization to work in your country; visa sponsorship not available.

Responsibilities

• Own document processing, chunking, and embedding workflows.
• Combine semantic search with exact-term retrieval, then rerank results for relevance.
• Develop agent retrieval strategies that determine what to inspect and when there is enough evidence.
• Evaluate retrieval quality with metrics and tests, including citations accuracy.
• Make early architecture decisions and take ownership of ongoing impact.

Skills

Retrieval systems

Search relevance metrics

BM25

NDCG

Recall@k

Independent work

Job description

Role overview

Own the retrieval systems that find reliable evidence in large collections of confidential documents and provide it to language models used in professional work. You'll shape the full pipeline, from document processing and search to relevance checks and decisions about what an AI agent should read, with measurable accuracy as a central goal.

Responsibilities

• Build and maintain document parsing, chunking, and embedding workflows.
• Combine semantic search with exact-term retrieval, then rerank results for relevance.
• Develop agent retrieval strategies that determine what to inspect and when there is enough evidence.
• Evaluate retrieval quality with metrics and tests, including whether citations point to the right passages.
• Make early architecture decisions and take ownership of their ongoing impact.

Requirements

• Strong experience building retrieval or search systems for large document collections.
• Knowledge of methods for evaluating search relevance, such as BM25, NDCG, or recall@k.
• Ability to build search abstractions that can work across multiple backends.
• Comfort working independently where priorities and processes may evolve.
• Existing authorization to work in the country where you are based; visa sponsorship is not available.

Nice to have

• Experience with agentic retrieval and assessing the quality of evidence passed to language models.

Benefits and work setup

• Full-time, fully remote work from eligible locations in the EEA, including Norway; local employment arrangements depend on location.
• Flexible, English-speaking environment with the opportunity to influence product and team practices as they develop.