Graduate In Training Program (GIT), Agentic AI Systems

Key Responsibilities

As a Graduate Internship Trainee in the VITE R&I Team, you will be actively involved in:

• Design and Architect Agentic Workflows: Move beyond simple prompting to implement complex reasoning patterns such as ReAct (Reason + Act), Plan-and-Solve, and Self-Reflection loops.
• Multi-Agent Orchestration: Develop and manage systems where multiple specialized agents collaborate (e.g., a "Dispatcher" agent coordinating with "Diagnostic" and "Navigation" agents) to solve complex user requests.
• Tool & API Integration: Build and integrate custom "tools" that allow agents to interact with external environments, including automotive APIs, real-time databases, and web services.
• Full-Stack POC Development: Develop lightweight, responsive web-based interfaces (e.g., Streamlit, FastAPI, or React) to serve as the front-end for agentic operational demonstrators.
• Agentic Evaluation & Guardrails: Implement testing frameworks to evaluate agent trajectory, tool-calling accuracy, and "loop-breaking" to ensure reliable performance in non-deterministic environments.
• System Prompt Engineering: Optimize complex system instructions and few-shot examples to steer agent behavior, memory management, and persona consistency.
  
  

Qualifications

Required Skills & Qualifications

• Academic Status: Must be a 2024- 2025 Graduate.
• Educational Background: Bachelor's degree in Mechatronics, Computer Engineering, Computer Science or a closely related field.
• Language Proficiency: Excellent command of both English and Arabic (written and verbal communication is essential). French is a strong plus.
• Agentic Frameworks: Proficiency in orchestration libraries such as LangChain, CrewAI, AutoGen.
• Backend & API Development: Strong Python skills with experience in FastAPI or Flask to create the "connective tissue" between the AI agent and the user interface.
• Prompt Engineering & Tool Calling: Deep understanding of structured output (JSON mode), function calling, and techniques for grounding agents using Retrieval-Augmented Generation (RAG).
• Frontend Basics: Ability to build functional UI/UX for demonstrators using modern web tools (e.g., Streamlit, Tailwind CSS, or basic React).
• Systems Thinking: The ability to map complex operational tasks into discrete agentic steps and handle edge cases where autonomous loops might fail.
• Modern AI Stack: Familiarity with vector databases and observability tools for tracking agent traces.
  
  

Program Details

Status: Full time.

Location: On-site.

Disclaimer: The successful candidate will be hired as a contractor and will be part of the VITE R&I Team at Valeo