AI Engineer Mistral / Prisme.AI

Project overview

GENERAL CONTEXT:

MISSION CONTEXT

In a dynamic of accelerating the availability of Generative AI solutions for its businesses and clients, the Group has entered into a major strategic partnership with Mistral.AI to provide a large-scale solution adapted to mass deployment use cases.

This partnership aims to provide a sovereign, secure, and controlled generative and agentic AI solution adapted to different operational contexts. To accelerate the deployment of the Mistral platform on internal infrastructure in a secure and scalable manner across the Group, and to deploy use cases and components leveraging Mistral's functionalities, the Gen AI program is looking for a qualified AI Engineer.

The AI Engineer works within one or more squads, under the responsibility of the AI Lead, to design, develop, and deploy generative and agentic AI solutions. He/she is responsible for the technical quality, security, performance, and adherence to deadlines for the implementation of assigned use cases.

Positioned at the interface between Product, Data, IT, and Security, the AI Engineer ensures alignment between business needs, technical constraints, and regulatory requirements.

MAIN MISSIONS

KEY SKILLS REQUIRED

SUCCESS INDICATORS

• Gen AI Program, based on two strategic partnerships: Prisme.AI (low code, small series use cases) and Mistral AI (large-scale platform, massive deployment), for sovereign, secure, and controlled generative and agentic AI.
• Deployment of the Mistral platform on the Group's internal infrastructure, within a complex and highly regulated banking IT system.
• Agile product organization with multiple squads. Involvement in one or more squads, under the responsibility of the AI Lead, at the interface of Product, Data, IT, and Security.
• 7 AI Engineers sought, to be integrated in several waves.
• Design and develop generative and agentic AI solutions leveraging the Mistral platform according to business specifications, ensuring their technical quality and regulatory compliance
• Integrate, configure, and optimize language models for mass deployment use cases (Mistral), evaluating the most suitable models and components for performance, cost, and compliance constraints
• Implement technical pipelines and set up quality control, monitoring, and hallucination/bias detection mechanisms in production
• Assist DevOps with the secure deployment of AI solutions on the Group's internal infrastructure and collaborate with Product, Data, IT, Security, and Compliance teams for business and regulatory alignment
• Document technical solutions, design choices, and best practices for developed solutions to ensure traceability and maintainability
• Participate in agile rituals and synchronization meetings to ensure consistency and alignment within squads and with the Gen AI program
• Share learnings, feedback, and innovations with Group teams to contribute to the evolution of AI standards and practices
• LLM
• Proven experience with large language models, including Mistral and other open-source and proprietary models
• Mastery of prompt engineering, few-shot learning, and prompt optimization techniques
• In-depth knowledge of fine-tuning, transfer learning, and adapting models to business contexts
• Understanding of attention mechanisms, tokenization, and LLM limitations (hallucinations, biases, drift)
• Generative & Agentic AI
• Experience in designing and developing autonomous AI agents and multi-agent systems
• Mastery of agent orchestration frameworks (LangChain, AutoGen, CrewAI, etc.)
• Ability to implement reasoning, planning, and decision-making loops
• Knowledge of agent-user interaction patterns and conversational context management
• Experience with the Mistral platform for large-scale deployment would be a plus
• Technical Stack & MLOPS
• Mastery of Python, with experience in agent orchestration frameworks
• Experience in model deployment and versioning
• Knowledge of CI/CD practices for models and containerization tools (Docker, Kubernetes)
• Ability to industrialize AI solutions in a real application context
• Close collaboration with IT and DevOPS teams
• Regulated Environment & Security
• Experience in a banking or highly regulated environment
• Mastery of GDPR, AI Act, security, explainability, and auditability constraints for AI systems
• Ability to implement safeguards and control mechanisms to limit risks
• Awareness of the ethical and responsibility issues of generative AI
• Soft Skills
• Intellectual curiosity and continuous learning ability in the face of the rapid evolution of LLMs
• Excellent technical communication and ability to explain complex concepts
• Close collaboration with product, data, and security teams
• Rigor, pragmatism, and ability to balance innovation and stability
• Quality and relevance of deployed generative and agentic AI solutions, measured by LLM evaluation metrics and end-user satisfaction
• Reduction of model hallucinations, biases, and drifts in production through implemented control and monitoring mechanisms
• Performance and latency of AI systems in real conditions, meeting defined response time and energy efficiency objectives
• Adherence to use case implementation deadlines and agile sprints, with regular delivery of features and improvements
• Adoption and engagement of business users and product teams with the developed solutions
• Contribution to accelerating the large-scale deployment of AI solutions within the Group via Mistral

Project details

• Start: flexible
• Remote: hybrid
• Location: Paris, France, FR
• Rate: 75-75 EUR/h
• Languages: French