AI Platform Architect – Cloud, MLOps & RAG

Project overview

Want to design the AI architectures of tomorrow? CELAD is waiting for you! 🚀

Do you enjoy designing complex architectures, industrializing AI platforms, and bridging the gap between Data, Machine Learning, Software, and Infrastructure teams?

We are looking for an AI Platform Architect to work with a player in the legal publishing and Legaltech sector, in La Défense.

You will contribute to the design and industrialization of an artificial intelligence platform using large documentary funds to offer advanced services to legal professionals.

🎯 Your Role

You will define the technical architecture of the AI platform and support its scaling, from cloud infrastructure to the deployment of models and inference services.

At the interface between ML, development, and DevOps/SRE teams, you will ensure the performance, availability, cost control, and security of the whole.

🚀 Your Main Missions

🛠️ Technical Environment

• Platform Architecture: Design robust, scalable, and secure cloud architectures to host AI services in production.
• AI Industrialization: Define MLOps chains, automate deployment pipelines, and structure the model lifecycle.
• Advanced RAG Architectures: Support the implementation of advanced RAG, Graph RAG, and agentic approaches.
• Infrastructure as Code: Industrialize the provisioning and configuration of environments with Terraform and Ansible.
• Orchestration and Deployment: Design containerized and orchestrated architectures with Docker and Kubernetes.
• Performance and Costs: Optimize CPU/GPU resources, inference, token consumption, and scaling.
• AI Response Quality: Contribute to improving retrieval, reranking, source traceability, and reducing hallucinations.
• Observability and Reliability: Define supervision, monitoring, and incident management mechanisms.
• Security and Compliance: Integrate requirements related to the confidentiality of legal data, GDPR, the AI Act, and data processing sovereignty.
• Transverse Leadership: Align architectural choices between ML, Software, Data, and Infrastructure teams, and support teams in their implementation.
• Cloud: AWS, Azure, and/or GCP.
• Infrastructure: Terraform, Ansible, Docker, Kubernetes.
• Development and AI: Python, PyTorch, TensorFlow.
• MLOps: MLflow, Kubeflow, Airflow, DVC.
• AI Serving: NVIDIA Triton and equivalent inference solutions.
• Data and Search: Spark, Elasticsearch, OpenSearch, Weaviate.
• Generative AI: LangChain, LlamaIndex, Advanced RAG, Graph RAG, agentic architectures.

Project details

• Start: flexible
• Remote: hybrid
• Location: 92800 Puteaux, France, FR
• Languages: French