Senior Researcher: Machine Intelligence - Theory of Large Language Models

Develop and lead an original research programme in machine learning theory, with potential emphasis on the foundations of modern neural networks and foundation models. Develop mathematical frameworks that explain learning, generalization, optimization, representation learning, architectural effects, scaling, or other important aspects of modern machine learning. Contribute to

the team's broader scientific direction by identifying theoretically important questions arising from empirical research and using theory to motivate new algorithms, architectures, or experiments. Disseminate research through peer-reviewed publications, conference presentations, open scientific engagement, and collaboration with the broader research community. Mentor interns and less experienced researchers, providing guidance and feedback on their research and development. A PhD in Machine Learning, Computer Science, Mathematics, Statistics, or a related field. Deep expertise and a solid record of original research in one or more areas of machine learning theory, such as learning theory, optimization and stochastic dynamics, high-dimensional probability and statistics, theoretical computer science, or the mathematical foundations of deep learning. Demonstrated ability to identify important research questions and independently drive research projects. Solid written and verbal communication skills, with the ability to explain complex ideas to a broad scientific audience and collaborate effectively across theoretical and empirical research. A sustained record of original, peer-reviewed research at leading venues in machine learning or theoretical computer science, such as COLT, NeurIPS, ICML, ICLR, STOC, FOCS, or SODA. Research that develops theoretical explanations for the behaviour of neural networks, foundation models, or other modern machine learning systems. Experience connecting theoretical analysis with empirical phenomena, experiments, or the design of learning algorithms and model architectures. Evidence of developing an original research agenda, leading research projects or collaborations, and mentoring less experienced researchers.