Software Engineering applicants have rated the interview process at Twitch with 3.3 out of 5 (where 5 is the highest level of difficulty) and assessed their interview experience as 75% positive. To compare, the company-average is 43.7% positive. This is according to Glassdoor user ratings.
Common stages of the interview process at Twitch as a Software Engineering according to 4 Glassdoor interviews include:
Skills test: 20%
Presentation: 20%
One on one interview: 20%
Background check: 20%
Phone interview: 20%
Here are the most commonly searched roles for interview reports -
I applied through college or university. I interviewed at Twitch
Interview
Spoke with a Twitch recruiter at a college career fair and was promptly contacted about a week later. Went through two phone screens and a 3 round onsite interview. All the interviewers were extremely friendly, willing to answer questions about the Twitch culture and work-life balance, and asked reasonable questions that you would expect from a standard interview.
Interview questions [1]
Question 1
Determine if tree is BST (with followups), Ice Cave, another third problem about DP that I couldn't remember the specifics of.
I did an online assessment, for which I obtained the maximum amount of points, but I was ghosted after and never heard back. I sent a follow-up email, but this was ignored as well. I was probably rejected because I'm a European applying for a NA role during the pandemic, but I would've appreciated a formal rejection.
Interview questions [1]
Question 1
Python data structures and algorithmic thinking. The questions were very original and related to Twitch.
I applied through other source. I interviewed at Twitch
Interview
Phone screen in which I discussed my past work and interests in Twitch and eSports in general, followed by a technical phone screen, followed by a series of in person interviews.
Interview questions [1]
Question 1
Imagine you want to design a service which outputs the total number of all concurrent viewers on Twitch to some output device every few minutes. You have a source of this information given as a data stream in which all currently active channels report their current number of viewers, which triggers every 10 minutes (your service must update more often than this).
How would you design this service?