Tubi Interview Questions

Tubi Interview Questions

Four of the 14 questions on this page are rated hard, a steeper skew than most company pages this size, and every one of them came out of a Tubi Data Scientist technical screen. Experimentation is the subject throughout: what actually drives a sample size calculation, how to spot and repair selection bias when treatment and control were never comparable to begin with, and what to say when an experiment reports several metrics that disagree with each other. Causal inference arrives in a form specific to ad-supported streaming, measuring the lift from a Super Bowl spot when there was no holdout and the whole country saw the ad, so difference-in-differences or a synthetic control has to stand in for randomisation. Two questions are pure statistics, including explaining to a product manager what a p-value does and does not mean when she wants a yes or a no. Only one is a data manipulation task, reconstructing daily net users from raw event logs where sign-ups and churn arrive as separate rows, and that is also the one you can run in the console. Written solutions cover five of the six, and most of the page reads without premium.

14 Questions 1 Company02.16.2026
Showing 14 results

Explore more Tubi interview questions

Jump straight to Tubi questions for a specific role or category.

In-depth guides