Cerebras (CA) logo

Cerebras (CA) Interview Questions
& Process

Real candidates share what happened, how many rounds they had,
and how the experience turned out.

Based on 31 interview experiences · FREE TO READ

3 Rounds average
Average Typical difficulty
51.6% Positive experience

Which role are you interviewing for?

8 roles · 31 reports

Candidate interview experiences

First-hand accounts from people who interviewed at Cerebras (CA).

Showing 4 of 31
Cerebras (CA) logo
Cerebras (CA)

Kernel Developer

Engineering · half a year ago

Mid Average Positive experience Accept offer 4 rounds
Interview process
Technical screen Technical screen Technical screen Technical screen
Interview formats
Coding Technical

It was a 4 round interview process. In each round, we were asked coding questions. The general coding questions were pretty similar to medium level LeetCode problems. We were also asked about parallel programming and matrix multiplication.

Confirmed questions4 questions
  • Can you implement 2D prefix sum?
  • Can you implement a multithreaded version of 2D prefix sum?
  • Tell me about parallel programming.
Cerebras (CA) logo
Cerebras (CA)

Machine Learning QA Engineer

Engineering · a year ago

Mid Average Negative experience No offer 4 rounds
Interview process
Recruiter call Phone screen Technical screen Onsite Onsite
Interview formats
Behavioral Technical Coding

A recruiter reached out to me about five months after I applied. First, there was a screening call with the hiring manager, where we mostly talked about my past experience and how it fit with the team’s needs. Next, I had a technical coding interview with a junior engineer, which went fine. I appreciated that the coding interview focused on implementing a real-world use case relevant to the job, not just a generic LeetCode problem. After that, I did a virtual onsite that included two more rounds. The third round was with someone who seemed unsure how to evaluate candidates, as the conversation was unfocused and they didn't really engage with my answers. They even asked who did my previous interview, which felt weird and out of place. In the fourth round, I talked to a director who seemed bored, kept yawning, and asked random questions without much interest in what I was saying.

Confirmed questions1 question
  • Can you explain the architecture of large language models?
Cerebras (CA) logo
Cerebras (CA)

Debug Engineer

Engineering · a year ago

Mid Average Positive experience Accept offer 5 rounds
Interview process
Recruiter call Phone screen Technical screen Onsite
Interview formats
Behavioral Technical

The interview process was pretty simple. First, I had two quick calls with a recruiter to make sure I was still interested and that my skills matched. Then, there were three more interviews: one with the manager, one with a team member, and one with a senior engineer. It felt pretty natural, moving from casual conversation to applying technical engineering skills.

Confirmed questions1 question
  • What's your strategy for tackling an unfamiliar problem in a component you haven't used before?
Cerebras (CA) logo
Cerebras (CA)

ML Engineer

Engineering · a year ago

Mid Average Negative experience No offer 2 rounds
Interview process
Recruiter call Technical screen Take home
Interview formats
Behavioral Technical Coding

Had an initial chat with HR, then a technical interview with the manager that lasted over an hour. The manager was friendly. They asked about my previous work and some ML/data science stuff. Then there was a live coding test, a standard linked list problem. After that, I heard nothing back. Reached out to HR and the interviewer but no reply. Not very professional.

Confirmed questions2 questions
  • Could you implement a basic linked list?
  • Can you code a simple linked list data structure?

Cerebras (CA) Interview Questions

Quoted word for word from Cerebras (CA) interview reports.

Can you explain how to calculate the prefix sum for a 2D array?

Read reports

Can you implement a multithreaded version of 2D prefix sum?

Read reports

Can you describe the GPU instruction pipeline and its differences compared to the CPU's?

Read reports

Can you explain the architecture of large language models?

Read report

Can you solve Leetcode 636. Exclusive Time of Functions?

Read report

Can you explain why std::unordered_map is often not the optimal choice for hash maps in many situations, and how TBB or ABSL's unordered hash maps offer advantages, specifying the ways and the conditions under which they are superior?

Read report

Formats, difficulty and experience

Across all 31 Cerebras (CA) interview reports.

Interview formats

Technical 40.6%
Coding 27.5%
Behavioral 27.5%
System Design 2.9%
Presentation 1.4%

Interview difficulty

Easy 6.5%
Average 90.3%
Difficult 3.2%

Candidate experience

Negative 32.3%
Positive 51.6%
Neutral 16.1%

Reports by job function

Engineering 31