Mobile Dev Interview In The Age of GenAI
How has the interview game changed since the release of GenAI tools?
Gergely Orosz and Elin Nilsson have recently published an interesting article about how tech hiring has changed since the release of GenAI tools:
They go over a lot of things in the article, such as how cover letters and resumes are now generated by AI which floods application pools, how interviewees nowadays “fake it”/cheat during coding interviews (algorithm puzzles) using GenAI tools, and much more. Read the article yourself, it’s a good insight into the current trends.
But as a mobile engineer prepping for interviews today what does that mean to you? What if you’re currently hiring mobile engineers, how do you adjust?
The rest of the article is my commentary on the matter, from both candidate and the hiring manager perspective. The TLDR:
cover letters were a low signal even before GenAI, very few people would actually craft theirs, even more so now. As an applicant, perhaps a way to standout is to actually write something unique yourself. As an interviewer consider dropping that field as it’s a low signal now even more than before. Contradictory, I know :)
resumes are not standardized (which makes it hard for the hiring manager to go through them fast) and were low value even before GenAI. LinkedIn could be the best solution here for both candidates and hiring managers.
Algorithmic puzzle style interviews were a low signal of a quality engineer even before GenAI, I’m glad that they are dying out and becoming irrelevant. Good riddance!
take-home assignments and follow-up pair programming sessions going over the applicant’s code are the best alternative signal to find quality engineers! I’ve been advocating for it for years!!!
system design interviews and/or behavioral interviews might get more weight now in hiring decisions.
Let me elaborate on each point a little more.
Cover Letters Are a Low Signal
Before GenAI and especially now with GenAI, cover letters effectively become irrelevant. Virtually everyone (or the majority) of applicants are crafting their cover letter specifically for the job description for the position, even if they are clearly not qualified. I can attest to it from my own personal experience as I’ve been hiring recently.
I myself would not spend much time on cover letters nor would I consider spending much time reading them when I’m hiring because they are not perceived as truthful anymore.
Ironically, and counterintuitively, one thing that might actually make you standout as an applicant nowadays IS to write your own bespoke cover letter that is not AI generated. If, let’s say, it’s way shorter and to the point, without mentioning every specific requirement in the job description, then there is a higher chance it would get noticed. But again, not every company is considering it a high enough signal.
For hiring managers, I advise either to make your cover letter requirement more open-ended (where applicants need to share their own opinion, which can’t be AI generated), or consider dropping your cover letter requirement for your applicants entirely. Why bother if it’s a low signal anyway?
Resumes Are Not Standardized / Lower Signal
As a hiring manager I find myself always, and I mean always, jumping to an applicant’s LinkedIn profile rather than trying to parse through unstandardized, sometimes poorly structured, resumes when I do a first pass on candidates. It is so much easier to look at the same standardized layout in LinkedIn where the title, job experience, and education are always all in the same place. This, of course, applies to the first phase of interview screening where the short list of candidates gets narrowed down.
Later though, after the shortlist gets identified, looking at resumes more closely sometimes helps, but also quite often all the same highlight details can be parsed out from LinkedIn much faster.
My advice for interviewees is to make your resume as short and simple and to the point as possible to make skimming through it easier. This is what recruiters and hiring managers do anyway. And please add a link to your LinkedIn somewhere at the top of the document!!!
For hiring managers, my best advice is not to get too hung up on the resume - they sometimes get “crafted” to perfectly match your job description just like cover letters do. Do quick skimming, look over LinkedIn - what you see in the resume should match what you see on LinkedIn as you can’t have it be crafted for each job application, see if the candidate has at least somewhat of a relevant experience. This should give you enough clues and confidence on whether to shortlist them or not, all other details you’d like to know about their experience can be learned throughout the interview process.
Algorithmic Puzzle Interviews Are Horrible And Less Relevant Now
Algorithmic puzzle style interviews were a low signal of a quality engineer even before GenAI, now it seems like they are becoming irrelevant due to candidates utilizing/faking/cheating this type of interview with GenAI.
I personally am glad that the “coding” algorithm puzzle style interviews are dying out and becoming irrelevant. It’s never been a good signal of how good a candidate is at software development. I’ve seen too many people who can crack puzzles well but can’t code and vise versa. Good riddance!
For interviewees, unfortunately, I recommend to still prep for this kind of interview. A lot of companies still attempt to use it as one of the interview steps, especially big ones that require conducting in-office interview where they can control the environment (whiteboard, popcorn, and the like, lol).
For hiring managers, I recommend just dropping the silly practice. Just drop it. It’s not worth it. None of your candidates will actually be solving algorithmic puzzles at their job after you hire them! There are better, much higher signal alternatives, such as take-home assignments and a pair programming follow-up interview.
Take Home Assignment + Pair Programming For The Win!
Pair programming has been the most effective way to simulate a real work scenario and to gauge how a candidate would actually perform at the job. Now, even more so in the age of GenAI.
Hiring managers, give your candidates a take-home assignment with a reasonable ask, such as to add a small feature to an existing sample codebase or something along those lines (something a candidate can finish within a few hours or a few days). Some of them will put their own effort into it, some will utilize GenAI tools, some will fully generate the whole solution with GenAI. At the end of the day, it doesn’t matter because you’d immediately see if the coded solution is too generic or simplistic or is not to your company’s/team’s standards. And even if it’s not as obvious for some reason, when you invite the candidate for a pair programming session interview, it will be immediately apparent if they understand and know the code they submitted. Those that made it themselves, with or without the aid of AI, would be able to clearly articulate what the code does and walk you through every nook and cranny of it. And those that did not will stumble and struggle.
The same goes for pair programming sessions with you, for example ask them to add a feature to the code that they wrote for their take-home assignment. You can even allow them to use GenAI if they really need to look up some stuff. You’ll be able to see how they operate in the real world solving real world problems along with you (in your sample take-home assignment codebase). This should give you a VERY strong signal on whether they are a good candidate to join your team or not.
System Design and Behavioral Interview Might Have More Weight Now
If the assumption that GenAI helps you solve low-level coding problems faster is correct, that means that software engineers will be solving coding problems at a higher level of abstraction than before. This means that system design skills will be even more important than before. Hiring managers would do well by putting more emphasis on this type of interview and the interviewees better be prepared for it even more than before.
Junior Devs Have The Biggest Challenge Now
The biggest challenge now is for junior developers to standout.
Like it wasn’t hard to get a job as a junior developer before, now it is even harder to standout due to the increased volumes of competition thanks to the flood of GenAI-aided job applications.
Pretty much the same advice as before applies to help junior developers standout:
have a short and concise resume, better LinkedIn profile
write a good cover letter (although, just like discussed above, perhaps this is not as applicable anymore)
showcase samples of your work and side projects (via Github and other means)
blog and otherwise share your experiences, however beginner it is
For hiring manager looking for junior developers:
look at their LinkedIn, just like with every other candidate, it can’t be GenAI generated for your specific job description
be on a lookout for candidates that blog and have side projects or code samples of their work on Github
give them take-home assignments
pair program to see how candidates perform in an actual work environment
Know Your Sh#!t and You’ll Be Just Fine!
All of these things rapidly changing make me even more resolute in my conviction that knowing your sh#!t is the way to go to get jobs. This is why I’m updating my upcoming 2nd edition of The iOS Interview Guide to help you level up yourself and be ready for any iOS interview!