Objectives of the internship

Artificial intelligence (AI) is one of the fastest growing areas in scientific research and R&D that will have a significant impact in almost sector of the industry. To ensure environmentally sustainable growth of AI, we need to consider the AI ecosystem, particularly energy consumption. To make the AI more sustainable, we should minimize AI energy consumption in end-to-end machine learning pipelines: data collection, model exploration, model training and runtime optimization and inference. 

The student will carry out his work at Nokia networks France in the Nokia Bell labs Network Service and Security Research (NSSR) lab of Bell labs Core Research.

During the internship, the intern will have to study the actual research on autonomous AI multi-agent system, sustainable AI as well as tools to measure the energy consumption of AI algorithms and models. The trainee will also develop a demonstrator for placing agents in nodes and locations with minimization of energy consumption and carbon footprint.

Artificial Intelligence, machine learning, multi-agents’ system

Programming: Bash/Shell script, Python

Systems/Software: Linux (very comfortable with), VirtualBox, Docker 

It would be nice if you also had: 

- Knowledge of GO language 

- Knowledges of Kubernetes 

The expected results are the following: 

Start of the art on autonomous AI agents orchestration frameworks, sustainable AI research in general (papers, frameworks, platforms)  A prototype-based demonstration and validation of placing agents in nodes and locations with minimization of energy consumption and carbon footprint. Measure and compare energy consumption of existing algorithms/models (ex. Deep Reinforcement Learning) Propose placement algorithms/models reduce energy consumption and carbon footprint.