Medior AI Engineer for a Legal Knowledge Platform

Lefebvre Sdu; Den Haag; Hybrid; EUR 75,000 - 110,000; Full time

Benefits offered by this job

End-of-year bonus

Profit sharing

Holiday allowance 8% and 12% personal/

Pension scheme

Travel and bike schemes

Hybrid work in The Hague

Growth opportunities

Job summary

Lefebvre Sdu invites a Medior AI Engineer to help shape the intelligent core of GenIA-L, our production platform used by lawyers and tax professionals across Europe.

You will build and maintain agents, connect knowledge packages, and evolve pipelines that keep our legal knowledge accurate and well-attributed. Collaboration with editors and subject-matter experts is central, and you’ll test ideas to sharpen the product while exploring deployment options.

Qualifications

• Several years of experience in software or ML engineering.
• Good knowledge of Python and experience with relevant machine learning libraries.
• Helped build an LLM application that runs in production.
• Hands-on experience with agent frameworks such as LangGraph and LangChain, tool calling, MCP or comparable orchestration.
• Measured the quality of an LLM system with evaluation sets or observability; MLflow and LiteLLM are a plus.
• Back-end experience, including building REST APIs.
• Experience in Kubernetes; Postgres is useful.

Responsibilities

• Build and improve GenIA-L agents and how they work together.
• Maintain knowledge packages with commentary, case law, parliamentary history and interpretation.
• Develop pipelines to create and update knowledge packages.
• Make quality measurable with evaluation sets and observability.

Skills

Python

Backend engineering

LLM evaluation

Independent work

Initiative

Tools

LangGraph

LangChain

MCP

Kubernetes

PostgreSQL

REST APIs

Job description

Lefebvre Sdu invites a Medior AI Engineer to help shape the intelligent core of GenIA-L, our production platform used by lawyers and tax professionals across Europe.

You will build and maintain agents, connect knowledge packages, and evolve pipelines that keep our legal knowledge accurate and well-attributed. Collaboration with editors and subject-matter experts is central, and you’ll test ideas to sharpen the product while exploring deployment options.