AI Data Exchange & MCP Intern (Java/EDC)

Company Overview

Amadeus is a global travel technology company that works to make travel better for everyone, everywhere by fostering innovation, building partnerships, and acting responsibly toward people, places, and the planet. The Nice office is exceptionally multicultural, with more than 4,500 people and 90+ nationalities working together in a collaborative environment.

Location

Nice, France

Mission & Objectives

The Model Context Protocol (MCP) is becoming the standard for connecting AI assistants to systems and data, but there is no standardized way to expose an MCP Server through a Dataspace Connector (EDC). As a Master’s Thesis intern, you will act as a generic bridge builder between the world of AI Agents and Sovereign Data Exchange. Your main goal is to provide a reference implementation for exposing MCP Agents via EDC.

Key Responsibilities

• State of the Art & Analysis: Deep dive into the Eclipse Dataspace Components (EDC) architecture (Control Plane, Data Plane, Transfer Process); master the Model Context Protocol (MCP) specification; analyze how an MCP connection (JSON‑RPC over WebSockets/Stdio) can be proxied or managed by an EDC Data Plane.
• Architecture & Design: Design a solution where an MCP Agent is registered as an EDC Asset; define how usage policies (ODRL) in EDC apply to an AI Agent (e.g., query limits, topic restrictions).
• Implementation: Develop an extension for the EDC (in Java) or a specialized Data Plane to handle MCP traffic; implement a mock MCP Server that wraps a sample dataset; ensure the “Agent” shares the same data context as the raw data product, ensuring consistency.
• Proof of Concept (PoC): Demonstrate a flow where a consumer (e.g., an LLM client outside the organization) negotiates a contract with the EDC to gain access to the internal MCP Agent.

Technology Stack

• Core Framework: Eclipse Dataspace Components (EDC), Java
• AI/Agent Standard: Model Context Protocol (MCP), JSON‑RPC
• Infrastructure: Docker/Podman, Kubernetes, Gradle
• Protocols: HTTP, WebSockets

What you will learn

• Sovereign Data Exchange: expertise on how companies share data without losing control (Gaia‑X, IDSA concepts).
• Agentic AI: work with the newest standards on how LLMs interact with external data.
• Open Source Contribution: opportunity to contribute findings or code back to the Eclipse foundation or the open‑source community.
• System Architecture: design high‑level interactions between distributed control planes.

Candidate Profile

• Final year student in Computer Science or Software Engineering (Master’s degree).
• Strong knowledge of Java and understanding of REST/HTTP APIs.
• Curiosity about Generative AI, LLMs, and the “Agentic” future.
• Ability to read complex technical documentation and specifications.

Diversity & Inclusion

Amadeus is an equal opportunity employer. All qualified applicants will receive consideration for employment without regard to gender, race, ethnicity, sexual orientation, age, beliefs, disability or any other characteristics protected by law.