I applied online. The process took 2 weeks. I interviewed at TikTok (Toronto, ON) in Nov 2023
Interview
Got an email for the Online Assessment. 5 Questions were asked from them 2 were multiple choice and other 3 were based on the DSA concepts coding. Their standard was high in considering the difficulty level.
Interview questions [1]
Question 1
The BytePlus platform has n servers, where the data stored on the ith server is represented by the array memory. To manage the server efficiently you can perform the following operation any (possibly zero) number of times:
• Choose an index idx, such that 1 ≤ idx s n/2.
• Swap the data of the pair of servers that are equidistant from the beginning and the ending of the array memory and have a distance less than or equal to idx.
The total working efficiency of the servers is calculated as the sum of the product of the data present in each server with the index of that server.
Given an integer n, and an array emory, find the maximum possible total working efficiency that we can get, since the total working efficiency can be very large print it modulo 109+7.
The interview process consisted of multiple rounds including an online assessment, technical phone screens, and virtual onsite interviews. The questions focused on machine learning fundamentals, coding challenges, and system design. The recruiters were responsive and kept me updated throughout the entire process.
Asked heavily about my resume, mainly focused on the projects I did that had overlapped areas with the team. Did an easy coding question on calculating a ML metric and asked about how to optimize the solution.
Interview questions [1]
Question 1
What was a metric I used in my recommender system project.
I applied online. I interviewed at TikTok (Singapore) in Feb 2026
Interview
The structure for both rounds is as follows:
20 mins deep dive into resume, explanation of personal projects & past experience, 20 mins testing about AIML concepts, 20 mins for leetcode medium
Interview questions [1]
Question 1
Explain what would happen if bias was removed from the equation during backpropogation
