RAG Engineer

MacroCoderz; Pakistan; Remote; PKR 1,200,000 - 2,400,000; Full time

Job summary

MacroCoderz is hiring a Senior remote RAG Engineer to design and own retrieval-augmented generation pipelines at scale. You will decide how documents are parsed, chunked, embedded and indexed, and how to combine keyword and vector search with effective reranking.

You will measure recall@k and MRRC performance, optimize latency, and contribute to cross-project standardization while supporting multi-tenant access. This role is fully remote worldwide.

Qualifications

• 4+ years of software engineering, including production search or retrieval systems.
• Hands-on experience with embeddings and at least one vector database.
• Understanding of BM25, dense retrieval, hybrid search and reranking.
• Experience measuring retrieval quality with recall@k and MRR.
• Strong Python proficiency.

Responsibilities

• Architect RAG pipelines: ingestion, parsing, chunking, embedding, indexing and retrieval.
• Implement hybrid search, metadata filtering, reranking, and multi-tenant access control.
• Build retrieval and answer-quality evaluation, and run experiments.
• Optimise latency and cost for large document sets.
• Review the retrieval work of other engineers and set standards across projects.

Skills

Python

Retrieval systems

Embeddings

Vector databases

BM25 & dense retrieval

Recall@k & MRR

Software engineering

Tools

pgvector

Qdrant

Weaviate

Pinecone

Elasticsearch/OpenSearch

Job description

Senior remote RAG Engineer at MacroCoderz: design retrieval-augmented generation systems - chunking, hybrid search, reranking and evaluation - at scale.

Retrieval-augmented generation is the core of most AI systems we ship: an assistant is only as good as the passages it retrieves. As a RAG Engineer you own that retrieval layer for client systems with thousands to millions of documents.

You decide how documents are parsed and chunked, which embedding models and vector stores to use, when to combine keyword and vector search, and how to rerank. Then you prove each decision with retrieval and answer-quality metrics.

What you will do

• Architect RAG pipelines: ingestion, parsing, chunking, embedding, indexing and retrieval.
• Implement hybrid search, metadata filtering, reranking and multi-tenant access control.
• Build retrieval and answer-quality evaluation, and run experiments against it.
• Optimise latency and cost for large document sets.
• Review the retrieval work of other engineers and set standards across projects.

What we are looking for

• 4+ years of software engineering, including production search or retrieval systems.
• Deep hands-on experience with embeddings and at least one vector database.
• Understands BM25, dense retrieval, hybrid search and reranking, and when each wins.
• Experience measuring retrieval quality with metrics such as recall@k and MRR.
• Strong Python.

Nice to have

• Experience with pgvector, Qdrant, Weaviate, Pinecone, Elasticsearch or OpenSearch.
• Document parsing for PDFs, tables and scanned files.
• Knowledge graphs or GraphRAG.

Location

: fully remote, open to applicants worldwide.