Muhammad Zuhaib Zahid

Hi there!

I build scrapers that report what they did not get.

Most extraction jobs fail quietly. A run finishes, the dataset looks reasonable, and the gap only surfaces later when a number fails to add up. My actors report their own coverage on every run: how many records exist, how many came back, how many pages were walked, and the reason the run stopped, named explicitly rather than left to be inferred.

That reporting is the whole point. It means a result can be trusted without auditing it by hand, and it means you are only ever charged for data that actually arrived.

Currently working on review and catalogue data for e-commerce.