Master Thesis Internship - AI Reasearch Multi-Agent Reinforcement Learning

Why join Ericsson? We are a world leader in the rapidly changing environment of communications technology – by providing hardware, software, and services to enable the full value of connectivity. You’ll play a part in using your skills and creativity to push the boundaries of what’s possible. To build never-seen-before solutions to some of the world’s toughest problems. Join a team of like-minded innovators driven to go beyond the status quo to build what’s next.

At Ericsson, you can be an explorer, a change maker and a force for good.
Our purpose: To create connections that make the unimaginable possible.
Our vision: A world where limitless connectivity improves lives, redefines business and pioneers a sustainable future.
Our values: Our culture is built on over a century of courageous decisions, in a place where co-creation and collaboration are embedded in the walls and where our core values of professionalism, respect, perseverance and integrity shine through in everything we do.

About this opportunity

Multi-agent reinforcement learning (MARL) provides a framework for coordinating multiple agents that share a common objective, such as base stations in a wireless network optimizing system-wide performance. A fundamental challenge in MARL is credit assignment: when a global reward improves or a constraint is violated, it is unclear how much each agent’s individual actions contributed to that outcome. This ambiguity is particularly severe in systems with tightly coupled interactions, delayed feedback, or partial observability, and can lead to slow learning, unstable policies, and overly conservative behavior.

In constrained MARL, where agents must also respect strict safety or regulatory limits, the credit assignment problem becomes even more critical. Agents must not only learn to improve collective performance but also determine how to adjust their actions to satisfy constraints without direct feedback for each individual contribution.

This thesis investigates the credit assignment problem in cooperative constrained MARL, exploring centralized-critic and counterfactual approaches, including COMA, difference rewards, and value decomposition methods. The work combines theoretical analysis with empirical validation on benchmark environments and a telecom use case, focusing on optimizing cell shaping while respecting regulatory constraints and energy budgets. The goal is to develop more efficient, scalable, and interpretable credit assignment mechanisms that enable better coordination in complex multi-agent systems, with demonstrated applicability to real-world wireless network scenarios 

You’ll get the opportunity to

• Explore state-of-the-art multi-agent reinforcement learning and constrained RL methods.


• Study centralized training with decentralized execution (CTDE) and counterfactual credit assignment techniques.


• Implement and analyze actor-critic and Lagrangian-based MARL algorithms in Python.


• Critically assess the limitations of existing multi-agent and constrained RL studies, and propose novel extensions.


• Evaluate the proposed methods in cooperative multi-agent simulation environments, potentially motivated by wireless communication systems.


• Contribute to research that may lead to a scientific publication.

Key skills

• You should be in the final year of a master's program (level 2) or engineering studies (level M2).


• You demonstrate theoretical knowledge and appetence for AI techniques and Reinforcement Learning


• You also have good programming skills, especially in Python.


• Additionally, familiarity with telecommunications is considered beneficial.


• You are fluent in English.

What we offer?At Ericsson, you´ll have an outstanding opportunity. The chance to use your skills and imagination to push the boundaries of what´s possible. To build solutions never seen before to some of the world’s toughest problems. You´ll be challenged, but you won’t be alone. You´ll be joining a team of diverse innovators, all driven to go beyond the status quo to craft what comes next.

Interns who join us will enjoy an outstanding chance to make connections, to make change, to make a real difference. To put it simply, we change how people, businesses and societies connect – and the results are genuinely exciting, challenging and often, awe-inspiring.

 
What happens once you apply?Click Here to find all you need to know about what our typical hiring process looks like.Encouraging a diverse and inclusive organization is core to our values at Ericsson, that's why we champion it in everything we do. We truly believe that by collaborating with people with different experiences we drive innovation, which is essential for our future growth. We encourage people from all backgrounds to apply and realize their full potential as part of our Ericsson team. Ericsson is proud to be an Equal Opportunity Employer. learn more.

Primary country and city: France (FR) || Massy

Req ID: 780760