Please stop the coding challenges

I have literally 10 years of experience in Android development (starting with Java and transitioning to Kotlin around seven years ago). There are many topics in Java and Kotlin that are irrelevant to Android development, but I focus only on what I need. I quickly learn and build on new topics when required.

I’m speaking here as a developer who has done more than just bug fixes or UI tweaks. I’ve built apps that have served millions of users. I even have personal projects that currently serve 20–30k users daily, with over 1 million downloads and a 4.9-star rating from 41k reviews. Beyond Android, I have experience managing servers with Nginx, setting up SSL certificates, using Cloudflare, AWS, and Hetzner.

I’ve designed an app with users, posts, and likes, handling both the front-end and back-end using cloud functions and Firebase. From writing Firebase rules to encrypting everything, I understood and implemented it all within a few months. I’ve also written numerous web scrapers in Python for personal projects, which are still in use. While some backend concepts might seem basic, I believe knowing how the backend works is a huge advantage for an Android developer.

Despite all this experience, I recently failed to solve a "number of islands" problem in an interview and got rejected:

Given an m x n 2D binary grid representing a map of '1's (land) and '0's (water), return the number of islands. This left me feeling useless. It made me question whether I’m fit to be a software developer. I studied DSA in college, but now I struggle to recall or prepare for it. Does this mean I’m a poor developer? Does someone who grinds LeetCode just necessarily have better skills in general? Is this what companies want? Are they okay to spend money and time in training someone just grinds LeetCode for the purpose interviews. Does someone who does this just for interviews is likely to stay in same company?

I believe there must be a better way to evaluate candidates. It makes sense to test DSA in campus interviews, where students are fresh and studying it. But for experienced developers, especially for roles like Android development, DSA is rarely relevant to the job requirements.

I’m not saying I can’t practice and do DSA—I can. But it has never been a job requirement for me, and I doubt it ever will be for an Android developer or a similar role. If it’s ever truly needed, I’m confident that most experienced developers, including myself, can excel through research and applied knowledge.

#Kill-DSA

I feel your pain. I've got similar amount of experience in my own area, I've been part of several successful software projects, and I have several popular open-source projects that I'm maintaining, but it all seemed to not matter a whole lot as I slogged month after month through interviews with rejection after rejection.

All I can say is that perhaps all those places that reject you are not actually the places where you really would like to work.

At one point I managed to get a job at a company which from the outside looked like a really cool place to work in. However all these interviews painted a very different picture of the work that I was actually assigned to do. I felt like I was being paid a senior engineer salary to do the job of a junior engineer. I was utterly bored and I complained that it's too boring and way below my skill level. This ended up getting me fired. Strangely, I've never had so many work-mates expressing to me how sad they are that I'm leaving. Several of them left the company not long after me.

Currently I'm in a happier place. Not getting as big of a salary but having an opportunity to work on a way more interesting software. And how did I land this job... after a single short interview they told me: would you like to start tomorrow?

I kind of agree with you that in practice very few jobs actually require any implementation of textbook algorithms.

The weird thing is that while people keep complaining about the focus on algorithms in the interview context, it seems like nobody objects when it comes to the recommended CS education.

I think there is a deep dissonance when people strongly recommend learning data structures and algorithms as part of CS and software engineering curriculum, but then when employers actually ask about these in interviews there is strong pushback.

I personally tend to think the right balance is in the middle. At least, whatever is taught in a typical software engineering curriculum should be fair game for interview questions.

The “number of islands” problem could be something that comes up in a DSA assignment in college.

I don’t know whether not knowing how to code it up in an interview setting makes you a bad engineer, but those who still remember their college classes got ahead of you apparently.

I often argue and get downvoted in the discussions when I claim algorithms and advanced math isn’t really needed as a prerequisite in the software industry. It’s only when people get gatekept for forgetting things taught in their college classes that they complain. Imagine that.

You’re not alone with this story, other people have shared similar stories. There are even a few very famous cases, like when the Homebrew maintainer complained about being rejected by Google https://news.ycombinator.com/item?id=15713801.

That said, you are making some assumptions. They might be right, but don’t forget to consider the possibility that they’re not, and then think about how to validate your assumptions, and about what it means and how to respond or adjust if your assumptions aren’t right. Also don’t forget to think impersonally about whether the job is the right fit for you. The job with the islands question might be a job for which it will be important to know and write similar algorithms, and based on the experience you describe, it might not be the right job for you. You sound like a full-stack dev, or even maybe a devops or generalist kind of person, and any given job might be looking for a more specific algorithm developer. When the job isn’t right for you, it is not a reflection of your existing talents or abilities, it could mean nothing more than they’re looking for someone with different experience or interest than you have. Being rejected might be best for you, before you know it.

> DSA is rarely relevant to the job requirements

I don’t know what this means. Generally speaking, a basic working knowledge of data structures and algorithms apples to damn near everything in computer programming jobs. I’d agree that most jobs don’t require knowing the subtle differences between parallel mergesort and parallel shell sort, for example. But most jobs that involve programming do require knowing some data structures and some algorithms.

Furthermore, companies are usually looking for well-rounded candidates. If I can hire someone who does full-stack Android and knows a lot of data structures and algorithms, that is better for me than hiring someone who’s great a full-stack stuff but gets stumped if they have to fix a connected components algorithm. Companies want devs who are good at as many things as possible, even when they say they’re looking for something specific.

Keep in mind that you’re competing with others, not being evaluated by yourself on your own merits. Also keep in mind that a single job rejection is nothing in the big picture, many things can go wrong in an interview, so look at the big picture and not individual interview questions.