I applied through an employee referral. I interviewed at TikTok (Mountain View, CA) in Aug 2022
Interview
Round 1 - ML Theory Fundamentals and Knowledge + case study for fraud/risk related field + experience
Round 2 - coding/algorithm or system design, ML Theory Knowledge + experience
Round 3 - system design + open question
Interview questions [1]
Question 1
What is the trade off between low and large learning rate?
What is the problem with high correlated features of collinearity?
If a precision is increasing but a recall is increasing how would the Precision Recall curve look like?
What are the disadvantages of undersampling in an imbalanced dataset? How to overcome it?
What is Precision ? What does it measure?
What is Recall?
Why do we need an activation function? What would happen if we remove activation function?
How to identify "bad" users?
A former colleague referred me to the Data Scientist position at TikTok, which kicked off the interview process. The technical screen included a question about designing an A/B test for a ranking model, followed by an SQL challenge related to user sessions. The funny part was that the SQL problem was a twist on one I had been grinding through on PracHub just a few days prior. After some behavioral interviews, I received an offer, but ultimately decided to decline as it didn't align with my goals.
Interview questions [2]
Question 1
Walk me through how you'd design an A/B test for a new ranking model on the For You feed, and how you'd pick guardrail metrics.
Given a sessions table (user_id, session_start, session_end, country), write SQL to find the users whose average session length grew the most week-over-week.
I applied online. I interviewed at TikTok (London, England)
Interview
one medium level sql test and an open question
Since i'm not familiar with business context(though i actually spent time preparing AB test and causal inference), i didn't get what i should answer.
Interview questions [1]
Question 1
if tiktok shop is going to launch in Netherland, how to evaluate it before AB test
Data Scientist applicants have rated the interview process at TikTok with 3.2 out of 5 (where 5 is the highest level of difficulty) and assessed their interview experience as 29% positive. To compare, the company-average is 37.3% positive. This is according to Glassdoor user ratings.
Candidates applying for Data Scientist roles take an average of 20 days to get hired, when considering 117 user submitted interviews for this role. To compare, the hiring process at TikTok overall takes an average of 30 days.
Common stages of the interview process at TikTok as a Data Scientist according to 117 Glassdoor interviews include:
One on one interview: 34%
Skills test: 26%
Phone interview: 19%
Background check: 7%
Presentation: 5%
Other: 3%
Personality test: 2%
Group panel interview: 1%
Drug test: 1%
IQ intelligence test: 1%
Here are the most commonly searched roles for interview reports -