Why I Keep Failing Candidates During Google Interviews…

Alexander Nguyen
Level Up Coding
Published in
4 min readApr 13, 2023

--

They don’t meet the bar.

After joining Google in 2022, I got trained to interview candidates. I’ve had a chance to meet candidates and saw how much they struggled. I’m ready to share some of their common mistakes (that I’m legally allowed to share).

Not knowing their data structures is NOT the #1 reason why I fail candidates.

Before we dig in, let’s go through what the interview process looks like today.

Online Assessment

Photo by Clément Hélardot on Unsplash

If you’re a new grad, you’ll likely be given an online assessment to solve a coding problem in whatever language you’d like. How well you do on this screen won’t affect your odds of getting an offer in the later interview stages.

What this screen is testing for is one thing:

Does the candidate know how to code?

If the candidate can’t solve the problem, then they wouldn’t do well on the phone screen or onsite. If they pass, congrats! They’ll move on to interview in person.

Phone Screen

Photo by Jenny Ueberberg on Unsplash

The phone screen at Google will represent one of the onsite interviews. The candidate will join a Google voice chat and solve a coding on Google’s IDE with a Googler.

The candidate will be held to the same standards as an onsite interview. This is to assess if the candidate can really pass back-to-back onsite interviews. Some do, most don’t.

For those that don’t do well on the phone screen, here’s what they should know:

  1. They can be given a second chance. Ask the recruiter and explain why you deserve another shot or didn’t demonstrate all your coding skills.
  2. They’ll have to wait again for at least 12 months to reapply.

If they make it onsite, then the candidate is in for some serious interviews.

Onsite Interviews

Photo by Rajeshwar Bachu on Unsplash

Making it to Google on-sites is a rare achievement in itself. For the few candidates that make it this far, here’s what they should know

  • It’s well known the candidate knows how to code
  • They may be assessed different scenarios and situations for each interview

Onsite candidates will be given 4–5 back-to-back on-site interviews. Most will consist of coding and a behavioral (Googliness) interviews as well as a system design interview for more senior engineers. Googliness is hardly an issue as long as the candidate can thrive in a professional workplace.

From the few candidates that make it this far, even fewer receive an offer. For the ones that don’t, Google legally will not provide feedback to candidates. If Googliness wasn’t the issue, then something went wrong in the coding interview.

Here’s what often goes wrong in the coding interviews

Coding Interview Mistakes

Photo by Maranda Vandergriff on Unsplash

Not clarifying the problem

Candidates jump straight into the problem without asking additional questions the moment the prompt is shared with them. The interviewer may correct them. But if they don’t, the candidate won’t realize they’re solving the wrong problem.

Focusing on unimportant details

If the problem is about using binary search, then focus more on the larger parts and come back to the smaller details. Some spend half the interview thinking about null exceptions or wrong input and never get to actually solving the question.

Staying quiet

Candidates are going to get stuck and the interviewers know this. What the candidate needs to do is know what questions to ask, think out loud, share their thoughts and take the interviewers input. Most candidates are dead silent and it’s usually because nobody teaches them how to communicate during a coding interview.

Assuming their code works

When the code is written, everyone wants to celebrate and be over with the interview. But the code most likely has bugs in it. That’s fine as long as the candidate goes through the example test cases and dry runs their code against it. Skipping this step isn’t professional.

Blurting the Running time

The candidate will absolutely be asked the running time of their algorithm. Most will just say it’s O(n) or O(nlogn) without any explanation. The better ones are the ones that explain why. The best ones are the candidates that explain the running time of different parts of their code and then reason where the bottleneck is.

Closing Thoughts

Most candidates who interview at Google do know how to code and likely practice their data structures and algorithms. The biggest difference between those who get an offer and those who don’t is how well they communicate their coding process.

Before You Leave

  1. Byte-sized system design newsletter!
  2. Let’s get personal on Discord

--

--