AI Agent Builder / Generative AI Engineer

Project overview

As part of structuring our AI division, we are evolving our technical trajectory: we are not looking for a classic Data Scientist or ML Engineer, but an experienced AI Agent Builder / LLM Engineer.

Your mission? Design, deploy, govern, and industrialize autonomous AI agents to meet diverse business needs (Support, Finance, HR, Operations, Fraud/Risk), while maintaining a robust and highly scalable data infrastructure on GCP.

You will work at the intersection of Data, Tech, Product, and Business teams to move from prototypes to real production.

YOUR MISSIONS

Design & Industrialization of AI Agents (Core of the role)

Data Modeling, Analysis & Inference

Data Infrastructure & MLOps (GCP Stack)

AI Acculturation & Culture Animation

Profile Sought

Work Environment

Consulting firm specializing in IT and Digital professions, Andema is located in Paris, Lyon, and Lille. We offer tailor-made IT services, adapted to our clients' needs, to ensure the success of each project. We connect our clients with competent consultants, who share their values and are fully invested in their mission.

At Andema, we firmly believe that the success of our clients' projects relies on the professional and personal fulfillment of our employees. That's why we place people at the heart of our priorities.

Joining the Andema adventure means evolving in an environment based on transparency, collaboration, intrapreneurship, continuous improvement, and the pleasure of working together!

• Creation of business AI agents: Develop and orchestrate functional and multi-tool agents to address concrete business problems.
• Governance & Guardrails: Implement control, evaluation, security, and alignment mechanisms to ensure reliable responses in production.
• Administration of Agent/LLM Platforms: Manage and supervise the orchestration and monitoring ecosystem (e.g., LangSmith, Azure AI Foundry, Vertex AI, Bedrock).
• Deployment & MLOps: Ensure robust production deployment with an adapted MLOps architecture (CI/CD via GitHub Actions).
• Hybrid modeling: Address key issues (fraud prevention, scoring, conversion optimization) by combining GenAI/LLM and classic ML / Deep Learning according to cost/performance impact.
• Impact Analysis & Statistical Rigor: Analyze transactional flow with rigor (trends, anomalies, causality) and challenge models on real data before any deployment.
• Data Pipelines: Design and maintain reliable data pipelines on our GCP stack (BigQuery, Dataform, Pub/Sub) to feed agents and models.
• Performance & Scalability: Optimize queries, manage large data flows, and eliminate bottlenecks.
• Inspiration & Pedagogy: Co-build use cases with Product and popularize your achievements for non-technical stakeholders without sacrificing scientific rigor.
• Watch & Standards: Establish development standards (tests, linting, docs) and actively maintain technological watch on advances in LLMs and agent technology.
• Minimum 4 years of experience in AI Engineering, LLM Engineering, Data Science, MLOps.
• Proven production experience: You have already designed, orchestrated, and deployed AI/LLM agents as well as Machine Learning models that run in real life (not just notebooks).
• Agent Platform Proficiency: Experience in governance, guardrails, and agent management via at least one market framework/platform (LangSmith, Vertex AI, Azure AI Foundry, Bedrock, etc.).
• Languages & Frameworks: Expert mastery of Python and SQL. Knowledge of PyTorch.
• GenAI & Agentics: Mastery of prompt engineering, RAG architectures, agent frameworks (LangChain, LlamaIndex, AutoGen, CrewAI, etc.), and evaluation/guardrails (Guardrails AI, NeMo, etc.).
• Data & Cloud Stack: Good mastery of GCP (BigQuery, Dataform, Pub/Sub, Vertex AI).
• MLOps Practices: CI/CD (GitHub Actions), containerization, monitoring and observability of models/agents.
• Databases & Viz: Practice with SQL, NoSQL databases, and data visualization tools (Tableau, Power BI, Toucan Toco, etc.).
• Pedagogy & Business Acumen: Able to convince with facts, explain the imperfections of a model, and say "no" if a solution is not ready for production.
• Autonomy & Proactivity: Ability to frame a problem, estimate its potential ROI, and propose suitable architectures.
• Analytical & Synthesis Skills: Scientific rigor combined with a very pragmatic business vision.

Project details

• Start: ASAP
• Duration: 1 an
• Remote: hybrid
• Location: Lille, Hauts-de-France, FR
• Languages: French