I tutor a small number of young people (middle- or high-school age) who are interested in machine learning and computer science.
I use both borrowed and original material, some of which is available online:
- unfinished-micrograd, forked from micrograd, with essential code removed for students to implement.
- numpytorch (private), a minimalist approach to backpropagation using Tensors. Based on PyTorch, uses numpy arrays. Unlike
numpytorchcontends with the difficulties that only arise when dealing with operations on tensors rather than scalars. Available on request.
- tic_tac_toe, a Q-learning agent learns to play tic-tac-toe in 500 lines.
I am open to students of all ages, but prerequisite knowledge of Python and/or differential Calculus is required. If interested, please include “Tutoring” in the subject line of your email.