AI Engineer (H/F/NB)

Why join Keyrus today?

Keyrus is an international group of 2,800 consultants and experts present in 28 countries, driven by a strong conviction: it is not artificial intelligence that transforms businesses, but intelligence designed, structured and integrated into their operations.

For over 30 years, we have built the data foundations that allow intelligent systems to create value. We design the operating system of the intelligent enterprise, where data, AI and human expertise integrate at the heart of business processes to generate sustainable performance. At Keyrus, we bring operational intelligence to life.

Artificial intelligence does not replace humans. It allows them to focus on what no system can reproduce: understand, decide, design and innovate. Joining Keyrus means building your judgment, expertise and ability to solve complex problems.

Technology is an accelerator. Keyrus culture makes the difference. Operational excellence allows us to connect the two.

AI Engineer

The AI Engineer builds, orchestrates and deploys artificial intelligence in production: generative AI bricks (Anthropic, OpenAI, Mistral, Google Gemini, etc.), machine learning models, multi-step agents integrated into client systems. It is the technical backbone of the Keyrus AI framework, ensuring that what is deployed works, scales and remains cost-controlled: each brick built carries its own FinOps responsibility, from prototype to production.

What awaits you

• You design and deploy end-to-end agentic AI systems: from the language model (Anthropic, OpenAI, Mistral, Google Gemini, etc.) capable of reasoning to the autonomous agent capable of acting in real production environments, not in isolated prototypes.
• You architect multi-step agents that orchestrate tools, memory and decision-making: you do not simply connect a model to an API, you structure how AI acts within the client's information system.
• You integrate what you build within client system architectures, leveraging available connectors, APIs and protocols (CRM, ERP, MCP, webhooks).
• You master the cost of models (Anthropic, OpenAI, Mistral, Google Gemini, etc.), tokens and inference (AI FinOps), so that the value created scales, not just in demonstration.
• You document each agent you build (scope, capabilities, limitations, KPIs), so it remains exploitable beyond your presence on the mission.
• You iterate based on real usage feedback (pilots, metrics), rather than considering an agent complete after its first deployment.
• You support the upskilling of internal teams on these topics, in the role of change champion.

What you bring

The Architect of Intelligence foundation: indispensable qualities, present in the criteria below: curiosity, creativity, communication, professional conscientiousness, culture, ability to execute.

Experience and background

• Production experience building generative AI (Anthropic, OpenAI, Mistral, Google Gemini, etc.) or agentic bricks, not just prototypes or workshops, regardless of your starting point: consulting firm, software publisher, startup, or end client.
• Practice in system integration (API, CRM, ERP, webhooks, connectors) in real production environments.
• Technical trajectory oriented towards growing autonomy in deploying AI solutions, with ambition to evolve alongside the Forward Deployed Engineer and Foundations Architect.

Posture and soft skills

• Engineering rigor that goes beyond demonstration validation: you think production, cost and maintainability from design onwards.
• Curiosity about the business result produced by what you build, not just the technology itself.
• Ability to produce, document and transmit what you build, so that an agent survives your departure from the mission.
• Enablement posture: you enjoy upskilling others, not just delivering alone.

Essential technical skills

• Mastery of a preferred language for AI: Python as priority, Java or Node.js/TypeScript as alternative depending on client ecosystem.
• Design of multi-step agent architectures (prompts, tools, memory, guardrails).
• Use of agent orchestration frameworks (LangChain, LangGraph, LlamaIndex, CrewAI).
• Implementation of RAG and LLM evaluation (Anthropic, OpenAI, Mistral, Google Gemini, etc.).
• Use of vector databases (Pinecone, Milvus, Chroma, pgvector).
• Practical mastery of APIs and SDKs from major model providers (Anthropic, OpenAI, Mistral, Google Gemini, etc.), with function calling / tool use.
• Implementation of evaluation and testing frameworks for LLM outputs (Anthropic, OpenAI, Mistral, Google Gemini, etc.): evals, guardrails.
• System integration within client architectures (CRM, ERP, MCP, webhooks, connectors).
• Use of low-code/no-code orchestration tools (n8n, Make, Dust).
• Mastery of Data Engineering fundamentals: SQL, modelling, data platforms (Snowflake, BigQuery, Databricks).
• Mastery of AI FinOps levers: semantic cache, model tiering, prompt optimization.
• Proven writing of technical documentation (agents, pipelines, integration architectures) intended for other teams.
• Fluent professional English, in writing and speaking, essential regardless of experience level.

What drives us

Collective Intelligence

Collaboration between expertise, disciplines and geographies allows us to combine knowledge and provide more comprehensive answers to our clients' challenges.

Reliability

Our ability to deliver complex projects with rigour and high standards is one of the foundations of the trust relationship we build with our clients.

Pragmatism

We prioritise concrete impact and measurable value creation over abstract technology discourse.

Entrepreneurial Spirit

Curiosity, energy and autonomy are at the heart of our culture. They foster initiative, innovation and continuous development of our expertise.

About Keyrus

At Keyrus, we support organisations in their transition from experimental AI to industrialised AI, from isolated agents to orchestrated systems, and from analysis to execution. This expertise is what we call the Architecture of Intelligence.

Through our proprietary Human Orchestrated Model™ (HOM), we design intelligent environments based on three fundamental pillars: robust and reliable data and intelligence foundations; governance ensuring humans remain at the heart of decision-making; performance management mechanisms enabling continuous improvement.

With over 30 years of expertise and 2,800 employees spread across 28 countries, Keyrus supports companies well beyond their transformation.

Keyrus is listed on Euronext Growth Paris (ALKEY – ISIN Code: FR0004029411).