Skip to content
León Palafox

Teaching 教育

Principles over recipes.

Current courses, past materials, and the occasional guest lecture — always hands-on, always from the mathematics up.

A bridge leading toward a luminous city, over a foreground of astronomical diagrams

Current courses

Universidad Panamericana

Deep Learning

Master's in Data Science

From the single perceptron and the Universal Approximation Theorem to the state of the art.

  • Optimization: SGD, Adam, and backpropagation mechanics
  • Architectures: CNNs for vision, RNNs/LSTMs for sequences
  • Attention and the Transformer architecture (BERT, GPT)
  • Generative models: VAEs and GANs
Syllabus & materials

Universidad Panamericana

Machine Learning 2

Master's in Data Science

The second course in the ML track: advanced supervised and unsupervised methods, model selection, and applied modeling workflows.

Syllabus & materials

Colegio de Matemáticas Bourbaki

Deep Reinforcement Learning

Specialized advanced track

Agents that learn from interaction, balancing mathematical rigor with implementation.

  • Foundations: MDPs, Bellman equations, and Q-learning
  • Policy gradients: REINFORCE, actor-critic, and PPO
  • Deep RL: DQN refinements, double Q-learning, offline RL
  • Applications: custom environments in finance and robotics

Course archive

Interactive

Pictures of ideas that are hard to hold still on a whiteboard.

Uncertainty

The jumping rope

Regression as a rope pinned at your observations: it thrashes where the data leaves it free and settles as you add points. Every frame is a real draw from the posterior.

Open

Guest lecturing

Machine Learning for Planetary Sciences — University of Arizona, 2016. Guest lecturer on applying CNNs to HiRISE imagery for geological feature detection.