About Idun
Idun Group is an AI consulting firm based in Paris (15th arrondissement). We audit how large enterprises actually use AI, then implement the use cases the audit identifies. Our clients are large accounts in aerospace, luxury, cosmetics and real estate.
A few delivered cases
- A luxury group running 13 AI agents.
- At a software publisher, agents in production: 80% of supporting documents processed automatically under human validation, a turnaround cut from 2 days to about ten minutes, 3 weeks from prototype to production.
- At a large social housing provider, 50 critical use cases uncovered by the audit.
- More than forty AI diagnostics in a year.
The role
You implement at our clients the use cases the audit has identified. You start from an identified need, not from a blank page.
You build on the AI tools the client already runs, and you develop custom software when they fall short. And if a use case won't hold, you are free to tell the client so.
Missions last from a few weeks to several months. Each one puts you in a new context: another industry, another information system, other teams.
What you will do
- Understand the client's business need and translate it into a technical solution.
- Take over the use cases from the audit and bring them all the way to production.
- Build on the AI tools already in place at the client, and develop custom software when they are not enough.
- Design and build AI agents, RAG systems and LLM integrations inside the client's information system.
- Take over a POC or existing code, including another vendor's, and industrialize it.
- Put guardrails in place: human validation, agent security, data protection.
- Deploy our tools in the client's infrastructure, on-premise included.
- Set up engineering practices: CI/CD, automated tests, code review, documentation.
- Measure adoption and gains once in production.
- Train the client's teams and hand over at the end of the mission.
The top skill
- Coding and organizing your projects with AI: Claude Code, Codex or OpenCode, with your own instructions, skills and sub-agents.
- Solid development foundations, to get real performance gains out of it without losing control of code quality.
Technical skills
Required
- Advanced Python (FastAPI, Pydantic).
- LLM APIs (OpenAI, Anthropic, Mistral, open-source models), and choosing the model by quality, cost and latency.
- RAG: hybrid search, chunking, embeddings, vector databases, evaluating answer quality.
- Agents: orchestration with LangGraph or LangChain, MCP.
- Docker, CI/CD, at least one cloud (GCP, Azure or AWS).
- Data and integration with the information system: SQL databases, REST APIs, data pipelines.
Nice to have
- Enterprise AI tools: Microsoft 365 Copilot and Copilot Studio, ChatGPT Enterprise, Gemini, Claude, Dust, Power Automate.
- Kubernetes and on-premise deployment.
- Google ADK, multi-agent systems.
- Monitoring agents in production: traces (Langfuse, OpenTelemetry), inference cost.
- Enterprise security: SSO, access rights, sensitive data, GDPR, EU AI Act.
- TypeScript.
- Sovereign AI: SLMs, Hugging Face, quantization.
- Model fine-tuning.
Consulting skills
- Running a workshop with business teams and coming out with a clear need.
- Scoping a perimeter, estimating effort, ruling out a use case that won't hold.
- Explaining a technical choice to a non-technical audience.
- Writing documentation the client's team can pick up on its own.
Profile
- Master's degree: engineering school or computer science.
- At least 5 years of software development.
- Hands-on experience with LLM or agent projects in companies: several projects taken to production, with real users.
- Able to act as technical lead on some missions.
- Regular watch on models, tools and AI practices, which you share with the team.
- Experience at an IT services company, a consulting firm or in a multi-client setting.
- Fluent French, technical English.
- Authorization to work in France required.
Permanent contract or freelance
- Permanent contract (cadre status), €55–70k gross per year depending on experience.
- Freelance: €550 per day excl. VAT, invoices paid without delay, through your own company or via umbrella employment (portage salarial).
- Location: based in Paris, on-site at least 3 days a week, at our 15th-arrondissement office or at the client.
- No full remote: fully remote applications will not be considered.
Why join us
- Today's technologies, used at clients: Claude Code, Codex, MCP, agents, RAG.
- An AI-native team. Code, deliverables and sales follow-up all go through Claude, with our own skills and our internal MCP server.
- A Claude license provided, and the chance to try new AI models and tools as soon as they ship.
- A small team, working directly with the founders.
How to apply
Send your CV to contact@idun-group.com with the subject "Senior AI Engineer IA générative". Include your availability, permanent contract or freelance, your expected daily rate or salary, and a GitHub link or a project in production.
Depending on the mission, an interview with the client is added to the hiring steps below.