Maksim Zubok is a doctoral candidate in Politics at Oxford University, Nuffield College. His dissertation explores ways of harnessing LLMs for social science research, from classic data labelling to using models as condensed snapshots of the internet to study how people organise relationships between concepts and form beliefs about the world.

He conceived and organised the inaugural school in 2023 and has led the programme through each edition since.

Joshua Berry is a DPhil candidate in Politics at Nuffield College, University of Oxford. His research combines insights from computational social science, behavioural economics, and networks to study political economy, regulatory enforcement, and judicial systems in advanced democratic countries.

More broadly, he is interested in developing experimental, machine learning, and computational methodologies for social science research.

Elodie Chervin is a DPhil candidate in Economics at Nuffield College, University of Oxford.

Elliott Ash is Associate Professor of Law, Economics, and Data Science at ETH Zurich's Center for Law & Economics. His research applies natural language processing and machine learning to large text corpora to study law, politics, and the economy.

His work uses text-as-data and large language models to analyse judicial decision-making, media, and political discourse, spanning computational social science, causal inference, and the economics of institutions. More information is available on his personal website.

Tatiana Shavrina is a Research Scientist Manager at Meta and has previously worked at Snap and AIRI. She is passionate about open source and multilingualism in LLMs, especially for under-resourced languages.

She contributed to BLOOM as lead for interpretability, led development of the mGPT multilingual model, and worked on Russian SuperGLUE and low-resource NLP methods. See her Google Scholar profile for publications.

Sergei Skvortsov is Lead Machine Learning Engineer at Nebius, focused on efficient training and inference for large language models. He regularly lectures on efficient inference at the Nebius Academy.

Previously, he led the inference team at Yandex Self Driving Group, where his team built the main inference engine powering neural models for autonomous cars and robots.

Ilya Boytsov is a member of the ML Performance Engineering team at Nebius, where he works on large-scale LLM inference systems and performance optimisation. His interests include efficient transformer architectures, distributed inference infrastructure, and scalable serving of foundation models, with a particular focus on Mixture-of-Experts architectures.

He co-organised the 2026 school alongside his teaching, and contributed substantially to the design of the collaborative research project.

Piotr Mazurek is Member of Technical Staff at Liquid.AI. He writes Tensor Economics, a newsletter on the economics and efficiency of large language model deployment.

At the school he lectured on the economics of LLM inference — how to reason about cost, latency, and deployment trade-offs when running models at scale.

Ibragim Badertdinov is a Lead Engineer at Nebius, working on data, code agents, RL environments, and evaluations. He is one of the co-authors of the SWE-rebench leaderboard — a live benchmark with fresh software-engineering tasks — as well as the SWE-rebench-v1/-v2 large-scale datasets for coding agents based on real-world tasks.

Previously, he worked on pre-training datasets. More information is available on his personal website. At the school he lectured on LLM data and RL environments.

Sagi Shaier is an AI researcher focused on foundational problems — efficiency, sparsity, reasoning, transfer learning, and continual learning — where progress ripples across the whole field rather than solving one narrow task. He draws on insights from biological systems to find simple, generalisable ideas, and cares about both the science and its real-world impact.

He completed his PhD in Computer Science at the University of Colorado Boulder in 2025, with a dissertation on factual knowledge-enhanced question answering in dynamic environments. His research and industry experience spans several U.S. national laboratories as well as Oracle, Cohere, and Aleph Alpha Research. He is currently a visiting professor at Johannes Gutenberg University Mainz. At the school he led a seminar, “Continual Learning: Why Models Forget, and What We Can Do About It.”

Margaret (Molly) E. Roberts is a Professor in the Department of Political Science at the University of California, San Diego, where she co-directs the China Data Lab. Her research sits at the intersection of new technologies and digital politics, with a focus on the politics of artificial intelligence, information control, and censorship.

She uses large collections of text, social media data, and machine learning to understand how technology reshapes access to information and political beliefs. Co-author of Text as Data: A New Framework for Machine Learning and the Social Sciences, her recent work explores the political neutrality of AI and how state media control and authoritarian censorship influence large language models.

Roberto-Rafael Maura-Rivero is a Research Engineer at Meta working on LLM agents and reasoning, and a Postdoctoral Researcher at the University of Oxford. He holds a PhD in Economics from the London School of Economics (LSE) and has held AI research roles at Google DeepMind and Amazon.

His research sits at the intersection of the social sciences and artificial intelligence, focusing on AI safety, alignment, and the economic impacts of transformative AI. He is particularly interested in applying social choice theory and game theory to democratise AI development, exploring how methods like reinforcement learning from human feedback (RLHF) can better align large language models with complex human values.

Christopher Barrie is an Assistant Professor of Sociology at New York University and a Research Fellow in the Department of Sociology at the University of Oxford. His research applies computational methods, natural language processing, and social media data to study political communication, protest dynamics, and how digital platforms shape behaviour.

His recent work engages closely with the methodological and social implications of generative AI, including reporting standards for the use of large language models in behavioural science, how political identity shapes human interactions with LLMs, and how generative AI might be designed to support democratic norms and reduce political polarisation.

Dr. Casas is an Associate Professor of Politics, Technology, and Computational Social Science at the Department of Politics and International Relations and the Oxford Internet Institute, and a Fellow at Reuben College. He is a UKRI Future Leaders Fellow and Director of the London Social Media Observatory. He received his PhD in Political Science from the University of Washington. Before joining Oxford, he held positions as Associate Professor at Royal Holloway University of London, Assistant Professor at Vrije Universiteit Amsterdam, and Moore Sloan Research Fellow at the Center for Data Science and the Center for Social Media and Politics at New York University.

Dr. Casas is a computational political scientist whose research examines how digital technologies and AI are reshaping politics, policy, and public discourse. In political communication and public policy, he studies the curation and moderation of political speech on social media platforms; the spread and effects of disinformation and information operations; the role of social media in shaping collective action and polarisation; and how social movements, interest groups, political parties, and the public use strategic communication to influence political agendas. In legislative politics, he investigates the conditions under which legislators shape policy through less visible mechanisms, such as amendments and legislative bundling. Across all his research, he develops and applies novel computational methods, including text and image analysis, large (visual) language models, and multimodal modeling approaches.

His work has been published in leading journals and outlets including Science Advances, American Political Science Review, American Journal of Political Science, Journal of Politics, Political Analysis, and Cambridge University Press. His research has received funding from NSF, NWO, ERC-Horizon, and UKRI, among other public and private organisations.