Group Data Insight and Innovation Senior Specialist

CMA CGM; Marseille; Sur place; EUR 85 000 - 110 000; Plein temps

Résumé du poste

CMA CGM is seeking an experienced AI/DevOps engineer to design, build and operate agentic AI solutions across the group. You will deliver end-to-end AI agents, RAG pipelines and tools, while enforcing engineering best practices, CI/CD and security.

You will industrialize deployments on AWS, containerize workloads with Docker/Kubernetes, manage environments as code and drive LLMOps for observability and cost monitoring. Strong collaboration with Engineering and Design Studio is essential.

Qualifications

• Master's degree in computer science or engineering.
• 5+ years of experience in software development and/or DevOps, including hands-on delivery of generative AI solutions to production
• Autonomous, delivery-minded and comfortable with exploratory, fast-moving topics
• Communicate fluently in English, both written and spoken
• Experience within CMA CGM ecosystem is a plus
• Mastery of AI-assisted development tools and their team configuration

Responsabilités

• Design and build end-to-end agentic AI solutions (agents, RAG pipelines, tools, connectors) for prioritized use cases.
• Deliver vibe-coded applications with AI coding agents and enforce engineering best practices (testing, code review, CI/CD, security).
• Define, monitor and improve the coding agent harness (rules, tools, guardrails, evaluations).
• Implement core agent components: orchestration, state management, conversational memory.
• Develop with Python and TypeScript using LangChain/LangGraph and major LLM APIs.
• Contribute reusable building blocks to CMA CGM’s AI platform.
• Design and maintain CI/CD pipelines for AI solutions.
• Containerize and deploy workloads on AWS with Docker and Kubernetes.
• Manage environments and Infrastructure as Code; ensure security/compliance.
• Operate in production with observability, tracing and cost monitoring.
• Document architectures and operational runbooks.
• Collaborate with Engineering and Design Studio teams from discovery to delivery.

Connaissances

Frontend React/TypeScript

Backend Node.js/TypeScript or Python

LLM orchestration LangChain/LangGraph

LLM APIs (Mistral/OpenAI/Anthropic)

AWS, Docker, Kubernetes

CI/CD & Infrastructure as Code

LLMOps & observability tooling

Git & collaborative development

English communication

Team collaboration & UX sensitivity

Formation

Master's degree in computer science or engineering

Outils

LangChain

LangGraph

OpenAI API

Mistral API

Anthropic API

Docker

Kubernetes

AWS

Description du poste

Job description:

MISSION

As part of the Group Data Department, within the Agentic Factory, you will design, build and operate the agentic AI solutions that transform how CMA CGM employees work: AI agents as well as applications built through AI-assisted development ("vibe coding") with solid engineering best practices. You will also define and monitor the coding agent harness that powers this delivery model. You will work directly with the Head of Agentic Factory and the Agentic Solutions Delivery Lead, and collaborate with the department's Engineering and Design Studio teams. Combining an AI developer profile with strong DevOps expertise, you take the Agentic Factory's AI solutions from prototype to production and keep them running reliably at Group scale. As a Group-level role, you will work hand in hand with the Group's GCCs and other entities (business lines, regions and subsidiaries) to roll out agentic AI solutions across the organization, in an international and multicultural environment.

WHAT ARE YOU GOING TO DO?

You design and build agentic AI solutions:

• Develop end-to-end agentic solutions (agents, RAG pipelines, tools and external connectors) for prioritized Group use cases
• Deliver vibe-coded applications with AI coding agents, enforcing engineering best practices (testing, code review, CI/CD, security)
• Define, monitor and continuously improve the coding agent harness (rules, tools, guardrails, evaluations)
• Implement core agent components: orchestration, state management, conversational memory
• Build with Python and TypeScript, using LLM orchestration frameworks (LangChain, LangGraph) and the APIs of the main LLM providers (Mistral, OpenAI, Anthropic)
• Contribute reusable building blocks to the Group's AI platform

You industrialize and deploy:

• Design and maintain CI/CD pipelines for AI solutions
• Containerize and deploy workloads on AWS with Docker and Kubernetes
• Manage environments and Infrastructure as Code
• Apply the Group's security and compliance standards

You run and operate in production:

• Implement LLMOps practices: observability, tracing and evaluation of agent quality
• Monitor reliability, performance and costs of deployed solutions
• Troubleshoot incidents and drive continuous improvement
• Document architectures and operational runbooks

You collaborate and raise the bar:

• Work with the Engineering and Design Studio teams from discovery to delivery
• Contribute to platform standards and engineering best practices
• Share knowledge with the team and keep a technology watch on the agentic AI ecosystem

Profile description:

Master's degree in computer science or engineering

• 5+ years of experience in software development and/or DevOps, including hands-on delivery of generative AI solutions to production
• Autonomous, delivery-mind and comfortable with exploratory, fast-moving topics
• Team player, open to feedback, with product and UX sensitivity
• You have a good command of:
  • Frontend development with React / TypeScript
  • Backend development with Node.js (TypeScript) or Python
  • LLM orchestration frameworks (LangChain, LangGraph)
  • APIs from the main LLMs (Mistral, OpenAI, Anthropic)
  • AWS, Docker and Kubernetes
  • CI/CD pipelines and Infrastructure as Code
  • LLMOps and observability tooling (logs, tracing, evaluation, cost monitoring)
  • Git and collaborative development practices
• You are able to communicate in English both written and spoken
• Previous experience within the CMA CGM ecosystem is a plus
• Mastery of AI-assisted development tools (Cursor, Claude Code, Copilot) and their configuration for team usage