Tell us about (or post links to) cool things you've built

Why do you want to work at Magic?✱

What is the largest dataset you've ever put together using unstructured or scraped data? What were the difficulties you faced in gathering and cleaning this data?✱

Even if you haven't built petabyte scale datasets, why do you want to do it and why do you feel you'd be able to get up to speed quickly (like within a week)?✱

What is a potential data source you think would be good for training large language models but isn't commonly used? Why do you think it's not being used?✱