Lead AI / Agentic AI Architect - Insurance Sector

Project overview

Mission Context

As part of accelerating its digital transformation, our client, a player in the insurance sector, wishes to strengthen its artificial intelligence capabilities to automate its business processes from end to end.

The objective is to design and industrialize autonomous and orchestrated AI agents capable of handling complex customer requests, particularly in the area of claims management, from their declaration to their closure, integrating the necessary human controls and validations.

The Lead AI will be involved in the entire lifecycle of the solutions, from scoping use cases to their deployment in production, with particular attention to performance, security, and return on investment.

Main Missions

Technical Environment

DomainTechnologies / SkillsDevelopmentPython, FastAPI, SQL, REST APIGenerative AI LLM, Prompt Engineering, RAG, Tool CallingAgentic Frameworks LangChain, LangGraph, CrewAI, AutoGenAI Models OpenAI, Mistral, Anthropic, open-source modelsVector Databases pgvector, FAISS, Pinecone, WeaviateCloud & Infrastructure Azure, AWS, GCP, Docker, KubernetesIndustrialization Git, CI/CD, MLflow, LLMOps, monitoringSecurity & Governance GDPR, AI Act, IAM, traceability, Human-in-the-loopBusiness Insurance Claims management, contracts, underwriting, compensation, customer relations

Candidate Profile

Expected Deliverables

Differentiating Criteria

Particular attention will be paid to candidates who can present concrete feedback from the insurance sector, especially on AI solutions that have enabled:

The main challenge of the position is to go beyond traditional documentary uses of AI to integrate intelligent agents directly into the core of insurance operational processes.

Work Environment

• Architecture and design of agentic AI solutions
• Define the technical and functional architecture of agentic AI solutions adapted to insurance processes.
• Design and orchestrate AI agents capable of handling customer requests end-to-end.
• Implement multi-agent architectures enabling the coordination of different business tasks.
• Integrate AI agents with existing information systems: CRM, contract management tools, claims platforms, and documentary repositories.
• Define mechanisms for supervision, traceability, error management, and escalation to business teams.
• Automation of insurance business processes
• Design intelligent automation solutions for the claims management cycle: declaration, qualification, collection of supporting documents, analysis, routing, monitoring, and closure.
• Develop AI agents for the classification and automated processing of customer requests.
• Implement mechanisms for extracting, analyzing, and controlling information from insurance documents.
• Automate interactions between AI agents and business applications via APIs.
• Integrate human validation points for sensitive decisions, particularly those related to compensation.
• Development and industrialization of AI agents
• Lead the development of generative AI and intelligent agent solutions in Python.
• Implement orchestration frameworks such as LangChain, LangGraph, AutoGen, or CrewAI.
• Design RAG architectures for the exploitation of internal documentary databases.
• Integrate different LLM models (OpenAI, Mistral, Anthropic, or open-source models).
• Deploy AI solutions in Cloud or On-Premise environments.
• Implement LLMOps / MLOps practices: versioning, testing, evaluation, monitoring, and CI/CD.
• Ensure the robustness, scalability, and security of deployed solutions.
• Governance, performance, and ROI measurement
• Define AI agent performance indicators: automation rate, accuracy, processing time, and quality of service.
• Measure the operational and financial gains associated with process automation.
• Identify risks related to hallucinations, processing errors, and data confidentiality.
• Implement human control and supervision mechanisms.
• Ensure compliance with GDPR, security requirements, and the regulatory framework applicable to AI.
• Contribute to defining the AI roadmap and prioritizing use cases based on their business value.
• Technical leadership and business support
• Act as the technical referent for generative and agentic AI topics.
• Supervise and support AI developers in the design and industrialization of solutions.
• Collaborate closely with Insurance business teams, IT architects, and Data teams.
• Lead scoping and use case design workshops.
• Provide feedback on AI projects deployed in the insurance sector.
• Higher education (Bac+5) in Computer Science, Data Science, Artificial Intelligence, or equivalent.
• Minimum 8 years of experience in Data / AI, with proven experience in designing and deploying generative AI solutions.
• Significant experience in the insurance sector, ideally on projects related to claims management or business process automation.
• Concrete experience in the end-to-end development and production deployment of AI agents, integrated into business information systems.
• Excellent command of Python and LLM / RAG / AI agent architectures.
• Ability to design robust, secure, and industrializable architectures.
• Experience in technical leadership and supporting development teams.
• Good understanding of the regulatory and confidentiality issues specific to the insurance sector.
• Target architecture for the agentic AI solution.
• Scoping and prioritization of insurance use cases.
• Developed, integrated, and production-deployed AI agents.
• Technical and functional documentation.
• Supervision, control, and evaluation mechanisms for agents.
• Performance and ROI monitoring dashboards.
• Industrialization recommendations and evolution roadmap.
• Partial or complete automation of claims management processes.
• Measurable reduction in customer request processing times.
• Improvement in manager productivity and quality of service.
• Integration of AI agents into complex IT environments.
• Demonstration of a measurable ROI after production deployment.

Project details

• Start: ASAP
• Duration: 6 mois renouvelable
• Remote: hybrid
• Location: Paris, France, FR
• Languages: French