Tech lead Expert Data Retrieval RAG - Freelance (M/F)

THE CONTEXT

Join a data adventure at the heart of Legaltech: our client is building the search and indexing engine that will provide access, tomorrow, to a massive and heterogeneous legal heritage for several teams across Europe.

Your mission is not to write a simple script: it is to design robust data pipelines, capable of ingesting, cleaning, and indexing enormous legal corpora, and then exposing fast, relevant, and hybrid (semantic + lexical) search at scale.

YOUR DAILY MISSIONS

1. Collection, ingestion & indexing

• Design ingestion pipelines from heterogeneous sources (publishing, open data, legal databases): parsing, cleaning, chunking, normalization.
• Choose and industrialize embedding models, feed and maintain vector databases.
• Guarantee index freshness, scalability on massive volumes, and traceability (lineage, versioning).

1. Search & relevance (Retrieval)

• Combine lexical and semantic search (BM25 + embeddings), reranking, and metadata filtering.
• Implement relevance metrics (recall, precision, nDCG, MRR) and maintain them over time.
• Structure the retrieval phase that feeds the LLM (RAG: context, citations, de-duplication, context window management).
• Instrument pipelines (logs, metrics, quality traces) and guarantee data security/compliance (GDPR, country-based partitioning).

1. Technical leadership & governance

• Promote a Clean Code, SOLID, and automated testing culture.
• Contribute to the group's technical influence (articles, meetups, open-source).

EDUCATION

Master's, PhD, or equivalent.

EXPECTED SKILLS

• Proven experience in data engineering / information retrieval on massive production volumes.
• Confirmed Retrieval expertise: indexing, embeddings, hybrid search - and the ability to measure then improve relevance.
• Strong Data culture: reproducible pipelines, data quality and lineage, dataset/index versioning.
• Solid AI / ML culture (embeddings, NLP, good understanding of LLM, without necessarily being the serving expert).
• A plus: appetite for other languages (notably Java) for cross-functional topics.
• Fluent English (daily European context).

SOFT SKILLS

• Product vision: your indexes and your search engine are managed as a product; other developers are your customers.
• Technical leadership & pedagogy: ability to evangelize best practices in data/retrieval/AI.
• Obsession with quality: absolute rigor regarding data security (Legaltech requirement) and service reliability.
• Innovation mindset: curiosity for advancements in AI, NLP, and data.
• Communication: ability to simplify complex concepts.

Python 3.11+ (mastery required), GoIndexingVector databases (pgvector, Weaviate, Qdrant, OpenSearch), hybrid search / BM25Embeddings & NLPsentence-transformers, embedding models, tokenization, document parsingRAG & LLMLangChain, RAG, reranking, OpenAI, Mistral, BedrockData & infraPostgreSQL, Airflow / Spark, Kafka, Redis, Docker, CI/CD, K8sTests & qualityPytest, relevance evaluation (recall, nDCG, MRR)