Tell HN: Don't work for companies that use hackerrank | Hacker News Reader

Tell HN: Don't work for companies that use hackerrank

24 pointsMCRed9 comments

This is not a way to evaluate programmers. This is taking the stupidity of the white board programming challenges and injecting it with steroids.

This is just another step down the path towards the commoditization and lower status of programmers.

Software development is an art, as well as a science, and a good engineer is able to architect systems, even small ones, and is able to comprehend the interaction of several modules of code.

Not bang out some crap that produces the correct results in a very narrow challenge.

Never mind that hackerrank is poorly implemented- it has numerous bugs (because it was clearly written by the kind of people who think hacker rank is a good idea- those who think speed is paramount over quality[1]) Numerous times it would not take submissions that were pasted into the editor (I used an external editor, Atom, because the syntax highlighting was correct for my language). It is limited in its languages and finicky in them. The errors are often incorrect, resulting in the rejection of working code. etc. etc.

But even if hacker rank were well implemented it is measuring only two things- speed and whether the output meets the criteria set by the creator of the test. This output criteria can often be wrong or more narrow than correct results.

And speed of banging out lines of code is not a measure of a programmer.

So, use this test and you'll get the people who don't understand software engineering, lack significant experience but can produce a lot of code quickly.

But not elegant systems... and you'll be slower to market and less nimble as a result.

[1] It isn't, not even in a startup. Speed to market is about getting a working product to market. This is slowed down by poorly engineered messes that don't work.

So, you don't do well on competitive programming problems and therefore decided to write a rant about it on HN. Do you even got an anectote where a candidate who did well on these kinds of problems were worse than another candidate who didn't do well on it?

My employer uses it for phone screens and it's nice to have a way to look at the same text and write/discuss some pseudo code. If someone is using it to time people on stupid questions then it's classic PEBKAC.

Is hackerrank measuring time or was that the interviewer?

DR Chrono has an add on the front page where people do their "challenge" in order to compete for an interview.

Worst example yet.

Apparently the company side generates scores based on whatever arbitrary factors HR chooses, including time, and some companies were simply comparing scores to decide who to hire.

Using it as a way of collaborating during a phone interview seems pretty reasonable.

Hi, one of the founders of drchrono here. I think the hackerrank challenge works well for us as a company and is better for most candidates than phone screening or resume reviews. It isn't perfect and it has strong and weak points like any interview process.

There are really great programmers we hired who first came through this hackerrank challenge that we would have missed if we had used more traditional interview methods like screening their resume and a phone interview. As an engineer myself I hate technical phone screenings and resume reviews, I HATE doing them when I'm hiring and I HATE being subjected to them if I was looking for a job.

Based on this post and some other feedback I've seen, there must be a set of good programmers (who companies like drchrono and others would love to hire) who do terribly on these types of tests or just don't want to commit to taking the 60 minutes or so to do these tests.

If you don't want to commit the time to do the tests and you want to blanket apply for tons of jobs at once, then larger companies/recruiters are a better fit and I'm ok as a tech startup if we miss out on you. We want to recruit people who are passionate about changing healthcare and our company's mission, and having someone commit an hour of time feels fair. We as a company spend 10+ hours for every candidate we screen so asking for one of your hours for 10+ of ours seems reasonable (I'm sorry if other people don't feel that way.)

We use Python/Django on our backend and we found the tests are reasonable to complete in Python. For iOS developers the C-type challenges seem an order of magnitude harder. So language experience and the technical limits of hackerrank do make this process have blind spots and we know we are going to miss out on some good candidates. I still think hackerrank is the best solution I've seen so far to objectively judge technical talent and works better than phone screens and resume reviews (which are horrible and I personally hate).

We've used hackerrank (formerly interviewstreet) almost since we started hiring software engineers, so every engineer we currently have working at drchrono passed through the hackerrank challenge as the first line interview. One other benefit is that if someone gets a good score on the test we feel 90% sure that the person is technically strong enough to be able to do their job and we focus the rest of our interactions with the candidate on other variables.

I think the best way to apply and get a job at a company that uses challenges like this if you are bad at them, is to have built awesome stuff: Have an app in an app store that people use, have a popular website, have contributed significantly to an amazing project/product) and to reach out through your network on Linkedin / Angellist to people at the company. This works better for candidates who have done a lot of stuff. I've seen programmers coming out of high school who have really cool apps in the App Store and have cool websites / web applications, so this is achievable for developers of any background.

The hackerrank site is terribly coded. Going between the problem and submission pages takes several seconds and hangs any browser I use about 25% of the time.

+1 to the incorrect solutions

It's obvious that it is not perfect and it should not be used to rank programmers during an interview but is a good test for basic algorithms skills, something that every programmer with a serious education should be mastering.

You basically code on the browser, which is not meant to be an IDE of your favorite programing language, you have to blame HTML, Web, Javascript, Web Browsers, Internet, your ISP, country, policies, cables, satelite all together, a lot of open source tools used by hackerrank for syntax highlighting, auto-completion etc.

Just because you didn't go through, you recommend whoever cleared not to join the prospective employer?

I see hackerrank and other similar sites are improving overtime, along with technologies & eco-system. You should understand that you might be programming 10000 miles away from the compiler.

I tinkered with HackerRank just for the coding exercises, and was initially impressed by the sheer volume of languages the site supports, but quickly grew frustrated with it. The use of stdio instead of a test suite like Codewars uses made writing in any languages for which they lacked I/O boilerplate tedious, and the layout was atrociously oversized, not even fitting completely in a 1000px-wide browser window.

Software development is an engineering discipline.

Reply on news.ycombinator.com