**1- Work Environment**

Internationally recognized generalist engineering school of the IMT (Institut Mines-Télécom), leading French engineering school (Technological University), IMT Atlantique aims to support transitions, train responsible engineers, and use scientific excellence to serve teaching, research, and innovation.

Within the STACK team (https://stack.inria.fr) of the Automatic Control, Production Systems, and Computer Science Department on the Nantes campus of IMT Atlantique, the position aims to contribute to the design, implementation, and validation of agentic services intended for data-driven applications deployed on Device-Edge-Cloud computing infrastructures. This work will be carried out as part of the national PEPR CLOUD project.

**2- MISSIONS**

Under the functional responsibility of the project manager and in collaboration with the project team, the missions are as follows:

1. **Design and develop agentic AI for network edge resource management**

2. **Evaluate and validate performance**

**3- ACTIVITIES:**

1. **Design and develop agentic AI for network edge resource management**

* Defines the agentic AI framework, including states, actions, and reward functions, tailored to network edge resource management and task offloading.

* Develops and implements Reinforcement Learning (RL) or Multi-Agent Reinforcement Learning (MARL) policies to autonomously make task offloading decisions on edge resources.

**2. Evaluate and validate performance**

* Evaluates the developed policies against baseline strategies, focusing on latency, throughput, energy consumption, CPU utilization, and system stability under varying workloads.

* Evaluates the adaptability and robustness of agentic policies in the face of dynamic workloads and changing network link conditions.

* Refines models based on experimental results.