Carrières

Your Mission

‍

Agentic architecture and orchestration

• Collaborative multi-agent systems design: definition of roles, exchange flows and decision-making logic.
• Master and implement low-code orchestration platforms such as AgentKit, n8n, Copilot Studio, etc.
• Model and orchestrate autonomous and supervised agents, managing routines, events, and interactions.
• Design architectures compatible with the Model Context Protocol (MCP).
• Implement mechanisms for traceability, observability and cost optimization (tokens, latency, performance).
  ‍

Cloud and applied artificial intelligence

• Work with AI environments such as OpenAI Platform, Azure AI, Azure AI, Azure AI, Azure AI, Google Vertex AI, Hugging Face, Anthropic Claude, etc.
• Interoperable design, hybrid and secure architectures, integrating several ecosystems.
• Use vector databases for semantic research and the management of contextual data.
• Manage the integration of models and external APIs (authentication, security, monitoring).
  ‍
  

Fine-tuning and customization of LLM models

• Perform fine-tunings on Hugging Face, OpenAI, Vertex AI or equivalent.
• Prepare and structure datasets (instruction-tuning, LoRa, LoRa, LoRa, PEFT, RLHF, synthetic data generation).
• Publish and deploy custom models on Hugging Face Hub, Azure ML, or GCP.
• Integrate these models into efficient multi-agent orchestrations.
  ‍
  

Programming and integration

• Develop flows and automations in node.js/TypeScript and Python.
• Use low-code orchestrators (n8n, Airia) to connect various platforms and APIs, etc.
• Work with existing REST APIs
• Integrate multimodal tools for automatic content generation.
  ‍
  

Prompt Engineering and Cognitive Design

• Create modular and traceable prompts, adapted to complex business contexts.
• Design and document instructions, sets, and agent roles
• Set up dynamic prompts: variables, contexts, safeguards, self-validation.
• Systematically document and iterate prompts to measure their performance.
  ‍
  

Data, observability, and MLOps

• Participate in the implementation of data ingestion and transformation pipelines (Snowflake, Synapse, BigQuery).
• Contribute to performance monitoring and model versioning (LLmops).
• Develop performance dashboards to track agents and models in production.