Python AI Developer

Python AI Developer (M/F)

Paris/Lille | Insurance Sector

Why this opportunity?

Because you will participate in building our artificial intelligence strategy, which we are currently defining.
You will join a motivated and talented team whose mission is to integrate AI at the heart of our business processes.

The goal is to meet the needs of our teams as well as our clients, and to automate certain low-value-added tasks.

Key elements

• Design RAG and agentic solutions applied to concrete business use cases.
• Develop and orchestrate agents while ensuring source traceability and controlling hallucinations.
• Directly contribute to the KELIA product configuration workflow, with human validation integrated at key stages.
• Participate in the production deployment of assistants, from their API to their integration into the internal web interface.

What you need to know:

We are looking for a Python AI Developer to join our R&D team in Paris or Lille.

You will be involved in the entire lifecycle of solutions, from design to production deployment, utilizing various approaches: language models (LLMs), supervised learning, clustering, and other machine learning techniques.

A significant part of the role also concerns product configuration: the objective is to extract relevant information from general terms and conditions, propose a configuration, ensure validation by an expert at each step, and produce a file directly importable into KELIA.

Main missions:

You will develop AI functionalities in Python.

• Document ingestion: extract text and tables from PDFs, Confluence pages, SharePoint documents, and GitLab repositories, then chunk, index, and update this content with asynchronous jobs.
• Search: hybrid search (semantic and keyword), filters by collection and path, reranking, and visual search on document pages.
• Agents and LLMs: develop agents with LangGraph (routing, search, filtering, reflection, response), write and version prompts, cite sources, and respond with "I cannot find the information" when it is missing.
• Product configuration workflow: contribute to the agent that extracts information from general terms and conditions, proposes the configuration, has each step validated by an expert, and produces a file importable into KELIA.
• Quality: maintain test sets, measure the correct answer rate, track fabricated responses, and verify that results are stable from one run to the next.
• Cost and latency: monitor token consumption, first response time, and cache, and choose the right model for each task.
• API: expose assistants via API, with streaming responses, and integrate them into the internal web interface.
• Document your code, write tests, and participate in code reviews.

This position is for you if:

• You hold a Master's degree (Bac+5) in Computer Science or Insurance.
• You are proficient in Python 3.11+ and comfortable with API development using FastAPI, Pydantic, and SSE streaming.
• You have experience or a strong interest in generative AI solutions, LLMs, and agentic architectures with LangChain, LangGraph, and LiteLLM.
• You are familiar with the principles of RAG, hybrid search, embeddings, and reranking, ideally with Qdrant.
• You know how to work with various document sources and their ingestion, especially PDFs, Confluence, SharePoint, or GitLab, using tools like Docling.
• You are comfortable with storage and persistence environments based on PostgreSQL and S3 object storage.
• You value the observability and quality of AI solutions, particularly prompt tracking, traces, logs, costs, and performance monitoring, using tools like MLflow.
• You are aware of security, authentication, and role management issues, especially with Keycloak.
• You are at ease working in a modern industrialization environment using Git, GitLab CI, Docker, Kubernetes, Helm, and pytest, as well as collaborative tools like Jira and Confluence.

What you bring

• A good command of Python and its best practices: testing (pytest), typing, dependency management.
• Your experience in using LLM APIs and the basics of prompt engineering, including structured outputs.
• Your understanding of how RAG works: embeddings, chunking, search, generation.
• A good command of SQL and database fundamentals.
• Git and teamwork skills.

Why join us:

• To contribute to a strategy where artificial intelligence is a key lever of the technological roadmap, with solutions integrated into the core business tools.
• To work on concrete use cases of generative AI, RAG, and LLM agents, from experimentation to industrialization.
• To combine cutting-edge AI technologies with strong business expertise in the insurance sector and create solutions with real impact.

• Ready to start a new experience?
  Send us your application and let's start the conversation

Find more English Speaking Jobs in France on Arbeitnow