§ 02 · Projects
Prototypes,
created and studied.
A small selection of ideas I've wanted to explore, and work that helps me in my own life.
Reinforcement Learning · Featured
Human-Aligned Policy Learning
Training RL agents that align with human preferences through reward modeling and imitation learning. Explored different approaches to gathering human feedback and train aligned policies.
Algorithmic Trading
Deep Learning for Pairs Trading
Deep learning methods for predicting structural breaks in mean-reverting equity pairs.
Machine Learning
What Makes a Hit Song?
We use the interpretability of classical machine learning to predict what attributes of a song are likely to make it a hit.
Machine Learning · Featured
A Transformer From Scratch: English -> Spanish Translation
Let's see how quickly a transformer can learn how to translate English sentences to Spanish. We'll dive into building a transformer, as described in the 'Attention is All You Need' paper, from first principles.