The scheduled interview was an hour long.

First 30 min. ML Design.

I was asked to design a recommendation algorithm for tiktok suggesting videos to users within their followed circle.
Discuss requirements. Metrics to optimize for and goal of the experiment.
Talked about collaborative filtering based model using matrix factorization. How will that scale given 1 billion users.
I suggested doing localized matrix factorization on similar users.
Talked about candidate selection a bit.
Offline and online performance evaluation ?
Baseline model to compare against.
How often to train the model and to decide ? Concept and data drift ?

Next 30 min. Coding

1. Reverse a linked list.
2. Followup: Reverse only given range in a linked list. https://leetcode.com/problems/reverse-linked-list-ii/