#data cleaning

5 data cleaning interview questions, asked at Microsoft, Nvidia, Perplexity and 2 more companies. Last asked Oct 2026.

Investigate bad data in a large clickstream export in a notebookCodingLockedYou work in a Jupyter notebook on your own machine, with an AI assistant allowed. First a short warm-up: sort names like item_2, item_10, item_1 so the number part compares as a…
Last asked Oct 2026
Fetch paginated records and filter them by rules and a historical averageCodingLockedAn HTTP endpoint serves patients' medical records, ten per page, with no page count given up front.
Last asked Oct 2026
Single-pass cleaning pipeline for a million crawled web documentsCodingLockedA 48-hour take-home: build a data cleaning pipeline. The input is a set of Parquet files with about a million documents crawled from public web pages.
Last asked Sep 2026
Clean a CSV of transactionsCodingWrite a small batch pipeline that cleans a CSV file of transactions. It reads the raw file, validates and normalizes every row, sets aside rows that cannot be fixed, and writes a…
Last asked May 2026
Clean OCR output into a training dataset for an LLMAI/ML designYou have a large amount of text produced by OCR from scanned documents: broken words, misread characters, page headers and footers mixed into the text, columns read in the wrong…
Last asked Feb 2026