.png)
The Cooperative AI PhD Fellows for 2025.
The Cooperative AI Foundation (CAIF) is proud to announce our first cohort of PhD fellows for 2025. We are delighted to welcome to our community these exceptional early career researchers who will contribute to advancing cooperative AI while receiving support through our fellowship program. We had 177 applications, with many high quality proposals that demonstrated a rapidly growing interest in the field of cooperative AI.
Our fellows were selected through a comprehensive two-stage evaluation process, incorporating assessments from both internal staff and external subject matter experts. The selection criteria focused on three areas: the potential impact of their research proposals, their demonstrated academic excellence, and their dedicated interest in addressing multi-agent and cooperation challenges in AI systems. The rigorous selection process underscores CAIF's commitment to promoting diverse perspectives in AI research. We look forward to the innovative contributions these fellows will make toward developing AI technologies that benefit society.
I work at the intersection of AI, Human-Computer Interaction, and social choice theory. I study the bidirectional relationship between AI and collective decision-making: developing AI systems to augment participatory processes like deliberative forums and, in turn, using these democratic approaches to make AI alignment more inclusive. I am particularly interested in how these methods can contribute to AI governance and ensuring AI agents fairly reflect diverse values and preferences across time.
Personal Website | Google Scholar
I am a Research Engineer at the Cooperative AI Foundation, where I conduct technical research on multi-agent systems. I am also a Foresight AI Safety Grant Recipient, working on multi-agent security, steganography, and AI control. Additionally, I consult with IQT's applied research and technology architect teams on AI, multi-agent systems, and AI infrastructure. Previously, I was a MATS scholar collaborating with Jesse Clifton on multi-agent systems research. In the past, I worked as an engineer for Dimagi on global health and COVID response projects. As AI systems become increasingly interconnected and autonomous, they introduce unprecedented risks that extend beyond the safety challenges of individual systems. My research focuses on understanding, measuring, and mitigating these risks via benchmarking, oversight, and security in both LM and MARL environments. Don't hesitate to reach out if you are interested in potential collaborations or want to chat!
Personal Website | Google Scholar | LinkedIn | X | GitHub
I am a PhD candidate at the University of Cape Town (UCT), deeply committed to advancing AI research and cultivating the growth of AI expertise in my home country, South Africa. With a strong academic foundation in Mathematics and Computer Science from UCT, I am passionate about nurturing local AI talent and contributing to the region's AI development. Under the guidance of Associate Professor Jonathan Shock (UCT) and Dr. Arnu Pretorius (InstaDeep Ltd.), my research delves into Offline Multi-Agent Reinforcement Learning (MARL). I focus on how agents can acquire cooperative skills from static datasets without requiring online interactions, addressing critical challenges for deploying AI in real-world contexts where live experimentation is often impractical.
Google Scholar | LinkedIn | GitHub
I am a second-year PhD student at NYU, advised by Prof. Eugene Vinitsky. I build agents that behave like humans to aid the development and evaluation of systems in multi-agent, safety-critical settings. To model human behavior, I combine reinforcement learning, generative modeling, and principles from cognitive science. Before my PhD, I completed an undergraduate degree in Neuroscience and a master’s in AI, where I studied cooperative game theory.
Personal Website | Google Scholar | X | GitHub
I am a Computer Science PhD student at Carnegie Mellon University, advised by Professor Vincent Conitzer, and part of the Foundations of Cooperative AI Lab (FOCAL). I strive to understand how to enable artificial intelligence and humans to effectively achieve better (social) outcomes in strategic interactions with other agents. More specifically, my research interests lie in algorithmic game theory and reinforcement learning, with an emphasis on cooperation and coordination of AI systems, trustworthy AI, and AI Safety. The tools I enjoy using include mathematical optimization, learning in games, computational complexity, and foundation models. Prior to CMU, I completed a master’s and bachelor’s degree in mathematics at Imperial College London and the Technical University Darmstadt respectively. In addition to that, I have also worked with the Fraunhofer-Gesellschaft (IEE) on machine learning methods for smarter renewable energy systems.
Personal Website | Google Scholar
I am a current master's student in machine learning at University College London, having graduated with a BA in maths from Cambridge in 2021. In the meantime I have worked on a number of projects in cooperative AI, including Welfare Diplomacy, a benchmark for language model cooperation, and RACCOON, an autocurriculum method for improving ad-hoc teamwork in cooperative multi-agent reinforcement learning. I am keen to continue researching ways to mitigate risks in multi-agent interactions, particularly when they involve open-ended, adaptive systems.
I am a first year PhD student in machine learning at the University of Oxford, supervised by Philip Torr and Christian Schroeder de Witt. I am currently focusing my research on multi-agent security and deception, and my research interests are in cooperative AI as well as in AI safety more broadly. My background is in economics and philosophy, so I have an additional interest in how my machine learning research can intersect with these fields.
I am an MSc student at the University of Amsterdam. Currently, I am an intern at the Krueger AI Safety Lab, working on Goal Misgeneralization and Unsupervised Environment Design supervised by David Krueger and Michael Dennis. Starting in January 2025, I will be a research intern at the Center for Human Compatible AI, working on preference learning, changing preferences, and influence incentives of assistive AI systems. I am broadly interested in topics at the intersection of Multi-Agent AI systems, Reinforcement Learning, AI alignment, Game Theory, and Cooperative AI. My research focuses on understanding and developing adaptive, robust, and safe goal-directed AI systems that collaborate effectively with humans and among themselves, both from a theoretical and empirical perspective.
Personal Website | Google Scholar | X | Bluesky
I am a first year computer science PhD student at the University of California, Berkeley. I work on problems at the intersection of AI safety and ethics and EconCS, and am supervised by Nika Haghtalab and Stuart Russell. My work in the past has focused on aggregating societal preferences, and on identifying failure modes of AI assistants. I am currently interested in developing the cooperative capabilities of AIs, especially to improve their ability to communicate, as well as in developing better AI assistants in general.
I am a PhD student at UC Berkeley advised by Sergey Levine. My research lies at the intersection of reinforcement learning and large language models, with a focus on how to facilitate safe interactions between AI and humans. My research includes building benchmarks for RL and LLMs, mitigating deception in AI systems, understanding the morality within LLMs, and building socially aware language models.
Personal Website | Google Scholar | GitHub
I am a computer science PhD student at UT Austin, where I am advised by Sriram Vishwananth. I am broadly interested in algorithmic game theory and cooperative AI. My research is often concerned with ways that strategic interactions involving advanced AI systems could differ from traditional game theoretic interactions (such as those among humans and corporations), and when and how these differences can be leveraged to improve the outcomes of these interactions. I was previously a summer research fellow at the Center on Long-Term Risk (CLR), a visiting scholar at the Foundations of Cooperative AI Lab (FOCAL) at CMU, and a scholar in the ML Alignment and Theory scholars (MATS) program. I received my bachelor's degree in math and computer science from Carleton College.
I am a PhD student at UC Berkeley advised by Stuart Russell and Sanjit Seshia. My research is focused on AI safety, particularly in the area of multiagent learning and human-AI interaction. The goal of my research is to inform the development of advanced AI systems so that they are beneficial and robust even when deployed in complex, multiagent scenarios. Previously, I received my bachelors at UT Austin where I did research at the intersection of formal methods, controls, and learning with Ufuk Topcu. I have also spent time at NASA Ames Research Center where I worked on planning under uncertainty.
Personal Website | Google Scholar | LinkedIn
I am an AI PhD student at the Center for Human-Compatible AI (CHAI) at the University of California, Berkeley. My research focuses on AI safety, using techniques like reward modeling, foundation model evaluations, and model interpretability to ensure that advanced AI systems act in a way that is beneficial for society. I have provided technical feedback for reports from the United Nations and World Economic Forum, as well as several books including The Alignment Problem by Brian Christian and AI: A Modern Approach by Stuart Russell. My other interests include cognitive neuroscience, computational social choice, creative writing and martial arts. Prior to joining CHAI, I studied at Duke, UNC and Oxford as a Robertson scholar, and worked at a fintech startup in London.
Personal Website | Google Scholar | X
I am a PhD student in the Computer Science Department at Carnegie Mellon University. Advised by Vincent Conitzer, I am a member of the Foundations of Cooperative AI Lab (FOCAL). My research involves problems in the intersection of theoretical computer science and artificial intelligence. Specifically, I am interested in using tools from algorithmic game theory and computational social choice to achieve cooperation in multi-agent settings, especially when the number of agents is ex ante unclear. Before CMU, I graduated summa cum laude from Harvard University, receiving a joint degree in Chemistry and Physics and Mathematics, with a secondary in Neuroscience, and a concurrent masters in Computer Science. While at Harvard, I was mentored by Ariel Procaccia. I am originally from Istanbul, Turkey. In my free time, I like cooking, hiking, and watching Turkish soap operas at 3x speed.
I am a Machine Learning PhD student at the University of Oxford, advised by Christian Schroeder de Witt and Philip Torr. My research focuses on reasoning, multi-agent systems, post-training, and AI safety. Before coming to Oxford, I was an undergraduate Computer Science student at UC Berkeley and worked as a Research Scientist Intern at MultiOn, focusing on agentic systems. My goal is to work towards safely developing the next frontier of AI systems, with an emphasis on RL training for LLMs in cooperative settings and inference-time search/verification strategies.
Personal Website | Google Scholar | LinkedIn | X
I am a PhD student at University of Maryland, College Park. I aim to build AI systems that align with human values and support human agency in complex settings. I have previously researched adversarial robustness of LLMs, watermarking LLMs and its complications, and algorithmic fairness. Prior to my PhD, I received my B.A. in Mathematics and Computer Science from Boston University.
January 15, 2025
