Senior RAG Engineer

Newcode.ai is a fast-growing legal tech and agentic AI company transforming how legal work is done. With teams across Norway, Sweden, Ireland, the US and growing, they work at the intersection of law, technology, and intelligence.

• At least five years building backend systems in production and at least two years shipping retrieval or RAG systems used by real users.
• Experience diagnosing and fixing retrieval regressions in production, including numerical evaluation before and after.
• Built and maintained golden query sets with labeling and trusted metrics influencing deployment decisions.
• Operated vector indexes in production (Qdrant), including parameter tuning, managing memory and latency, and reindexing without downtime.
• Strong Python skills with experience in FastAPI, PostgreSQL, and Redis under load.
• End-to-end ownership including maintenance and tech debt.
• Curiosity and tracking developments in vector databases and large language models.
• Experience with AI coding assistants and informed practices for reviewing AI-generated code.
• Comfortable working in a fast-changing startup environment.

Experience expected with stack:

• Python: 5+ years
• FastAPI: 2+ years
• Vector databases in production (Qdrant): 2+ years
• Embedding models and retrieval quality metrics
• PostgreSQL / SQL
• OCR-heavy document ingestion at scale; background in information retrieval methods like BM25, NDCG, recall@k; experience with search abstractions over multiple backends.

• Own the retrieval pipeline end to end: document parsing and chunking, embeddings, indexing, and maintenance as clients add or change material.
• Expose retrieval services through FastAPI, running ingestion and reindexing as background jobs.
• Design search combining semantic and keyword retrieval in Qdrant, with ranked list fusion, metadata filtering, and reranking.
• Build agentic retrieval including query decomposition, multi-step loops, and managing cost and latency budgets.
• Develop an evaluation layer with golden sets, retrieval metrics, realistic regression tests, and tracing to problematic document chunks.
• Ship fast, dependable services using Python and FastAPI handling heavy background jobs under load.
• Ensure strict data security and tenant isolation for privileged client material.
• Adapt retrieval to support multiple languages beyond English, handling compounding and inflection challenges.

Role is fully remote within the EU. Candidates must have the existing right to work in their country; no visa sponsorship is provided.