Senior applied research scientist at Mila. 7+ years of industry experience training and deploying deep-learning models.

Vim Tetris

less than 1 minute read

The classic game of tetris with vim keybindings

*Note: Disable vimium on this page if you are using it.

Python’s ‘shelve’ is really useful for LLM debugging

2 minute read

Python’s shelve module is a persistent key-value store perfect for LLM debugging.

1 minute read

Make arXiv papers easily available in LLM-friendly formats.

Using LLMs to directly generate the html for a response.

4 minute read

Comparing multimodal agents and how they struggle on ‘The Password Game’