Where machines learn to reason. Build, train and align the foundation models shaping the next decade.
Our School of Artificial Intelligence sits at the core of the university — the discipline every other school draws on. Students train large models on our own compute cluster and study how to make them capable, interpretable and safe.
From deep learning theory to deployed systems, you work alongside researchers publishing at NeurIPS and engineers shipping models to millions.
Admissions →Deep learning, optimization and the theory of generalization.
View department →Language models, reasoning and multilingual understanding.
View department →Perception, generative imaging and multimodal systems.
View department →Interpretability, control and robustness of frontier models.
View department →Agents that learn to act in complex environments.
View department →Deploying models to healthcare, science and industry.
View department →A solid grounding in linear algebra, probability and calculus helps. First-year courses bring everyone up to speed before the advanced material.
Yes — students train models on the AI Supercomputer Lab cluster from their second year, with allocated GPU hours for projects.
Both. The curriculum is anchored to research labs, and every student ships an applied capstone with an industry partner.
Applications for the 2026 intake are open. Scholarships available for international students.