Senior Software Engineer Python - AI

Project overview

The Context

Our client is a European player in legal tech. Within the AI Service team, you will be responsible for the raw material of AI: the collection, ingestion, indexing, and relevance search on massive legal volumes, the foundation that powers the group's language models and products.

Your playground: designing robust data pipelines and exposing fast and relevant search (semantic and hybrid) to serve multiple teams across Europe. You will be involved in making structural decisions about the stack.

The Missions

The Team

You will join a team of passionate, close-knit, and supportive experts who tackle the highest technical challenges in legal publishing. You will work with multiple teams across Europe.

Why This Role is Interesting

Company Culture

A long-established European family group, present in several European countries, driven by entrepreneurial and humanistic values and a long-term vision.

A culture of expertise and knowledge sharing, serving legal and tax professionals, with a commitment to reliable AI.

• Design the collection and processing of heterogeneous sources (publishing, open data, legal databases): parsing, cleaning, chunking, and normalization of documents
• Choose and industrialize embedding models, feed and maintain vector databases, manage index freshness and reconstruction
• Ensure the scalability of indexing on massive corpora and data traceability (lineage, index versioning)
• Define the retrieval strategy: lexical and semantic search (BM25 + embeddings), reranking, and filtering by metadata
• Implement test sets and relevance metrics (recall, precision, nDCG, MRR) to continuously measure and improve retrieval
• Structure the retrieval phase that feeds LLMs: context, citations, de-duplication, context window management
• Instrument pipelines (logs, metrics, quality traces) and ensure the security and compliance of legal data (GDPR, country-specific isolation)
• Apply and promote Clean Code, SOLID, and automated tests, and contribute to the group's technical influence (articles, meetups, open source)
• Get involved at the time of structural decisions, with real influence over the final stack
• A rare data challenge: indexing and making a massive and heterogeneous legal heritage relevant, in service of AI
• A long-term vision: create the group's industrial standard for the coming years, rather than delivering an isolated feature
• A close-knit and supportive team of experts, in a hybrid setting in Courbevoie (2 days remote work)

Project details

• Start: flexible
• Remote: hybrid
• Location: Paris, France, FR
• Languages: French