All courses

Linear Algebra

Vectors, matrix factorizations, and eigendecomposition from first principles — the mathematical foundation every other course in this catalog assumes.

Optimization

First-order methods and convex toolkit for training models — from gradient descent to Adam, with enough convexity to know what your optimizer can and cannot guarantee.

Reinforcement Learning

Policy gradient, Q-learning, and actor-critic methods built from the Bellman equation up. Train agents in MuJoCo and Atari.

Robot Learning

Imitation learning, behaviour cloning, and sim-to-real transfer. Build controllers that learn from demonstration and adapt to reality.

Physical AI & Vision-Language Models

From vision backbones to embodied agents — how language and perception fuse into systems that act in the physical world.

Generative Models

VAEs, diffusion models, and flow matching from first principles. Build systems that synthesize images, audio, and structured data.

Probability & Statistics

Probability, conditioning, and estimation for machine learning — from random variables and Bayes to MLE/MAP, uncertainty, and the language of modern generative and RL models.

Advanced Optimization

Boyd-depth duality and cone programs meet second-order, distributed, and sharpness-aware training — the bridge from convex structure to modern large-scale ML optimization.