Home

Giuseppe Paolo

I am a Research Scientist at Cognizant AI Lab, where I study how autonomous agents learn, coordinate, and develop complex behavior in shared, open-ended environments.

My research spans agentic and multi-agent AI, embodied intelligence, reinforcement learning, and open-ended, nature-inspired systems.

I received my PhD from Sorbonne University in 2021 and my MSc in Robotics, Systems and Control from ETH Zürich. I have also been a visiting researcher at HKUST and IBM Research Zürich.

Email • Google Scholar • GitHub • LinkedIn

Recent News

July 2026 - TerraLingua opened to the public. Anyone can now introduce an autonomous agent into the live ecology and observe how it interacts, adapts, and contributes to an evolving artificial society.

May 2026 - I was featured in Cognizant AI Lab’s Behind the Research series, where I discuss TerraLingua, multi-agent systems, embodied AI, and the importance of connecting research with real-world applications.

March 2026 - We released TerraLingua: Emergence and Analysis of Open-endedness in LLM Ecologies, a platform for studying cooperation, institutions, shared knowledge, and cumulative culture in populations of LLM agents.

November 2025 - We released Solving a Million-Step LLM Task with Zero Errors, introducing MAKER, which combines extreme task decomposition with multi-agent voting to execute very long processes reliably.

July 2025 - AdaPTS, our framework for adapting univariate foundation models to multivariate probabilistic forecasting, was presented at ICML 2025.

June 2025 - I joined Cognizant AI Lab as Research Scientist

April 2025 - Our paper Zero-shot Model-based Reinforcement Learning using Large Language Models was presented at ICLR 2025.

February 2025 - We released TAG: A Decentralized Framework for Multi-Agent Hierarchical Reinforcement Learning, together with its open-source implementation.