Lead AI Engineering

LineUP7 is a Paris-based consulting agency founded in 2015 specializing in customer experience optimization through data, AI, and digital marketing tools. We serve luxury, tourism, banking, retail, and real estate sectors. We are hiring a Lead AI Engineer (CDI, F/H) reporting directly to the COO to establish technical standards and accelerate our agentic AI practice, managing architecture choices, team skill development, and pre-sales contributions.

You know what separates a chat LLM from a real agentic platform in production. You'll own the design, industrialization, and operation of agentic AI systems for enterprise clients: business process automation, conversational agents, intelligent pipelines.

Key responsibilities:

• Define architectural standards and reusable AI assets across engagements
• Design hybrid AI architectures (cloud, on-premise, multi-model) adapted to client constraints
• Develop and industrialize RAG pipelines: chunking, embedding, reranking, ingestion, indexing, retrieval, quality evaluation
• Deploy and operate agents in production with guardrails and governance
• Drive orchestration stack choices and evolution for agentic systems
• LLMOps: monitoring, drift detection, versioning (MLflow, Weights & Biases)
• Work with model providers (OpenAI, Anthropic, Mistral, Hugging Face)
• Lead full development cycles from prototype to production
• Contribute to pre-sales and technical documentation

Requirements:

• 5+ years of engineering experience, 2+ in production AI/ML projects
• Mastery of agentic orchestration frameworks: LangChain, LangGraph, LlamaIndex, CrewAI or equivalents
• Strong Python, FastAPI/REST APIs
• Docker, Kubernetes, cloud (GCP, Azure, AWS)
• Knowledge of data platforms (Snowflake, Databricks, dbt) and CRM/marketing stacks (Salesforce Data Cloud, SFMC, Adobe Campaign) appreciated
• Architectural judgment under client pressure, technical documentation rigor, and build/buy/integrate decision-making

Contract: CDI (permanent), on-site in Paris 14e arrondissement.