Have coding interviews gotten harder?
Summary
TLDRThe transcript covers the evolution of coding interviews over time, contrasting the experience of a former Google tech lead in 2009 with the more rigorous algorithm-focused interviews seen nowadays on platforms like LeetCode. There is discussion around how the interview bar has risen as more engineers specifically prepare for these tests, with companies responding by asking increasingly difficult technical questions. The speaker ultimately argues this arms race is unhealthy, as mastery of algorithms and data structures is not the sole measure of a strong engineer. He does confirm though that Google still mostly asks LeetCode mediums rather than hards in interviews.
Takeaways
- 😊 Conrad didn't use LeetCode to prepare for interviews back in the day because it didn't exist yet
- 📚 Instead he used books like Cracking the Coding Interview and practiced with others
- ⏳ LeetCode was created in 2015, so it's a relatively new resource for interview prep
- 📈 Google initiated the trend of asking more algorithm and data structure problems in interviews
- 🚀 Many LeetCode questions come directly from Google's internal question bank
- 🤯 The bar for passing coding interviews has gotten higher over the years as more engineers specifically prepare
- 🙅♂️ Asking LeetCode hards in interviews often backfires by testing if candidates have seen that exact question before
- ☹️ The overemphasis on algorithms in interviews doesn't necessarily surface the best engineers
- 😅 You're still more likely to be asked a LeetCode medium than a LeetCode hard, even at Google
- 😡 Don't be a jerk interviewer who only asks LeetCode hards!
Q & A
When did the concept of the modern coding interview start, and who pioneered it?
-Google pioneered the modern coding interview with data structures and algorithms problems in the 2000s.
What resources were engineers using to prepare for interviews before platforms like LeetCode existed?
-Before LeetCode, engineers prepared by reading books like Cracking the Coding Interview and practicing problems together.
How do many LeetCode questions originate from Google's internal question bank?
-Many LeetCode questions come from Google's internal question bank, often being posted publicly years after first being asked in interviews.
Why does the interview bar seem to keep getting higher as engineers prepare more?
-As more engineers specifically prepare for algorithm questions, companies compensate by asking harder questions to maintain their hiring bar.
What are the issues with companies asking extremely difficult algorithm questions?
-Asking very hard questions often just tests if candidates have seen that exact problem before, which is not a good evaluator.
How has the state of the economy impacted coding interviews recently?
-In the current economy, the coding interview bar has been exaggerated and made more difficult.
What is the most common difficulty level of coding questions asked today?
-Most coding questions today are still at the medium difficulty level, even at top companies like Google.
Why are algorithms questions not the best indicator of engineering skill?
-Algorithms knowledge is important but not a complete indicator of what makes a good engineer.
What advice does the speaker give to interviewers?
-The speaker advises interviewers not to ask extremely hard algorithm questions.
What resources does the speaker recommend for preparing for interviews today?
-The speaker recommends LeetCode as better preparation for modern coding interviews than books like Cracking the Coding Interview.
Outlines
😂 Lead code interview practice in the old days
The paragraph discusses how software engineering interviews worked before platforms like LeetCode existed. The author's former lead at Google mentions solving around 400 LeetCode problems now, but when he was interviewing 10-15 years ago, candidates practiced with books like Cracking the Coding Interview and mock interviews. There were no standardized online judge platforms with curated algorithm questions back then.
😲 Google created the modern coding interview
The author explains how Google was the first company to start asking data structures and algorithms style coding problems in interviews. Many LeetCode questions are sourced from Google's question bank. The author interviewed at Google in 2009 and got a dynamic programming problem that later showed up on LeetCode years afterward. This demonstrates how over time, Google interview questions end up becoming public.
😔 The coding interview bar keeps getting higher
Due to candidates specifically preparing for coding interviews, companies have increased the difficulty over time to maintain signal. Now medium questions are common, and some companies even ask hard questions occasionally. The author argues this is problematic, as being good at algorithms does not define a good engineer. While hard questions are still uncommon, the general difficulty has increased over the years unfortunately.
Mindmap
Keywords
💡Coding interviews
💡LeetCode
💡Cracking the Coding Interview
💡Levels of difficulty
💡Supply and demand
💡Being a good engineer
💡Memorization
💡Current economy
💡Interviewers
Highlights
LeakCode was created in 2015, so it makes sense people weren't asking questions from it before then.
Cracking the Coding Interview was the main prep book before LeetCode existed.
Google started the modern coding interview with data structures and algorithms problems.
Many LeetCode questions come from Google's question bank.
A LeetCode question was asked in a Google interview years before it appeared on LeetCode.
Google doesn't ask questions that have been publicly leaked.
The coding interview bar is getting higher as engineers specifically prepare for these interviews.
If everyone can solve easy questions, companies will start asking mediums. If everyone can solve mediums, they'll ask hards.
Asking hard questions that require candidates to have seen that exact problem before is problematic.
With the economy, the coding interview bar today is definitely higher than years ago.
Algorithms and data structures don't determine if someone will be a good engineer.
Candidates today are still more likely to be asked a LeetCode medium than hard, even at Google.
Asking LeetCode hards in interviews is excessive.
LeakCode was created in 2015, so it makes sense people weren't asking questions from it before then.
The coding interview bar today is definitely higher than years ago.
Algorithms and data structures don't determine if someone will be a good engineer.
Transcripts
the other week I had a conversation with
my former Tech lead from Google here's a
short clip from it that I thought was
interesting uh how many leak code
problems have you solved if I go to my
profile right now it's probably around
400 so I guess that goes into the next
question when you were doing your
interviews coding interviews how were
you preparing yeah so when I was
interviewing it was a little bit
different I actually got an internship
at Microsoft that was my first job and
there was no lead code back then so back
then you would be reading cracking the
coding interview
uh you would be going through some
questions that you've heard other people
going through and you would you would
just practice together then the actual
interview was pretty similar to what
happens today in B Tech but after the
internship there was no other interview
um for a full-time conversion at
Microsoft at Google there was but I went
through the industry route which which
just meant like a little bit of a more
difficult interview um with a system
design as well so to give a little
context leak code was actually created
in 2015 so it makes sense that people
weren't really asking questions from it
back then whereas cracking the coding
interview is a book that was written in
2008 so that was kind of all you had
like people read books but not only that
if you actually go through cracking the
coding interview today you'll quickly
realize that it's just not enough to
prepare for the modern coding interview
because the interview is just so much
more difficult nowadays Conrad mentioned
that in 2009 at Google he had more of a
traditional leak code interview but in
case you didn't know Google is actually
the reason that we have the modern
coding interview they were the first
ones to start asking these data
structures and algorithms problems and
people don't realize it but a lot of
questions from leak code are actually
just taken from the Google question Bank
how do I know that well in 2009 I also
interviewed for Google and I got a
variation of Jump game a dynamic
programming or greedy problem after my
interview I went on leak code to check
if the problem was already on there but
it actually wasn't now a few years ago I
was doing aite code contest and all of a
sudden I saw this problem that I had
already solved before it was the
variation of Jump game from my Google
interview 3 years after I was asked the
question it was then posted delete code
but I will say now that the question has
been public for a while you're probably
not going to be asked it because Google
does not ask any questions that are
publicly leaked but back to the original
question it's starting to seem that
since Engineers are very specifically
preparing for these algorithm style
interviews that the bar is getting
higher and higher because the entire
concept is supply and demand companies
only need a certain amount of people if
everybody can solve leak code easys
companies are going to ask medium
questions if everybody can solve mediums
they're going to start asking hard
questions and that's when things start
to go wrong because most people can't
solve hard problems unless you've
already seen that problem before and
that's not a good signal are you just
trying to hire people that have already
seen your specific problem before I
think that's when things really start to
break down and it's really getting to a
point with the current economy that it's
kind of been exaggerated now I haven't
interviewed in a while but I think it's
fair to say that the bar today
definitely is higher than it was several
years ago and that's very unfortunate
because as much as I like algorithms and
data structures it is not the bottom
line of what makes a good engineer I
want to end by saying that you're still
more likely to be asked a medium
question than a hard question yes even
at Google but if you're an interviewer
watching this video and you ask Elite
code hards come on don't be a dick
関連動画をさらに表示
Is Grinding LeetCode Worth It?
Algorithms Explained for Beginners - How I Wish I Was Taught
How to Start Leetcode (as a beginner)
I solved 541 Leetcode problems. But you need only 150.
Confessions from a Big Tech Hiring Manager: Tips for Software Engineering Interviews
How to Solve ANY LeetCode Problem (Step-by-Step)
EX-Google Recruiter Reveals Secrets To Never Failing A Tech Interview
5.0 / 5 (0 votes)