Oct 03 2026
~ 9 min read
Well, unfortunately, after a large cost-cutting restructure at my current employer, I’m back on the other side of the interview table again, looking for my next role.
That means updating the ol’ CV, polishing my story bank (STAR-friendly, of course), and deciding where to invest time refreshing my technical knowledge and skills. More dauntingly, though, I’m practicing condensing years of relevant professional experience into a conversation that typically lasts less than an hour.
I fully understand why companies need a thoughtful hiring process. The stakes are high, and the costs of a bad hire are significant, especially when teams are lean and budgets are tight.
But after interviewing for roles over the past several months, that "thoughtfulness" hasn’t always been apparent. Some processes feel stuck in the past, disconnected from what it takes to build modern software, let alone in the AI-era.
I’ve had interviews that were really excellent: focused and grounded in real-life scenarios, while others have left me puzzled about what, exactly, the company was trying to assess in the first place 🤔.
So a question keeps coming back to me: Would I hire me today?
More specifically, could the people already doing these jobs pass the tests they expect new candidates to pass? And would the results actually tell us who could do the job?
What can coding tests actually tell us?
Entire businesses have grown around technical assessment, including platforms like Karat, HackerRank, and CodeSignal. Whatever tools we use, though, we need to be clear about what we want to learn.
Here’s my take: writing code is "basically" dead, but being able to understand code and recognize what solid code looks like, should be hotter than ever.
I say "basically" because engineers still need to know how to write code. Building something can tell us how someone breaks down a problem, tests assumptions, handles edge cases, and recognizes when their first approach isn’t working.
But producing code from memory against a stopwatch shouldn’t be the centerpiece of the evaluation. In the actual job, we have documentation, existing code, colleagues, and increasingly AI tools available to us. Being effective means knowing how to use those resources, while still taking responsibility for the result.
Being able to read, understand, and evaluate code deserves much more attention. Especially when it’s generated by AI. Knowing which patterns fit the problem is part of that judgment, along with checking whether the implementation handles failure and can be maintained.
Give me a bug to investigate in an existing system, for example, and you can see how I approach something unfamiliar. What do I ask first? How do I narrow down the problem? Can I explain what I think is happening, and what I’d check next? Finding the answer matters, but the path someone takes to get there tells you quite a bit about what it would be like to work with them.
I’d really like to see more assessments built around those skills, with less emphasis on LeetCode-style problems. If AI is part of the job, let candidates show how they use it and verify the result.
Evaluating Senior Leadership
Now for something that I believe should really be quite obvious: the depth of your evaluation should reflect the breadth of the role.
It’s here where I’ve personally experienced an even bigger head-scratcher: making a coding test mandatory for a “Head of” or “Director” level role.
I totally understand wanting engineering leaders to be technically credible. Architecture reviews make sense: evaluating high-impact decisions, questioning operational risks, and recognizing when a team needs support are all part of the job. A title shouldn’t give anyone a free pass.
But if I’m hiring someone to lead several engineering teams, the assessment should also tell me how they develop people, make difficult trade-offs, and connect technical investment to business priorities.
So what would I ask instead? Something like: your team is missing delivery commitments, production incidents are increasing, and Product wants to launch a major new feature. Where do you start?
I’d want to understand what they investigate first. Who do they involve? Is this a capacity problem, a technical problem, or a lack of clarity? What do they delegate, and where do they personally step in?
That conversation gives me much more to work with. We’re discussing technical judgment, people leadership, and business priorities together, which is pretty much how they show up in the job.
Hidden messages in the hiring process
As a candidate, pay attention to the types of assessments and the order in which they’re given. Technical rounds first? Ask how much hands-on technical work the role involves. People/cultural rounds first? Ask how those capabilities factor into the decision. Are the rounds “all or nothing”? I’d want to understand why a particular result outweighs the rest of the evidence.
The order can also reflect interviewer availability or an early screen intended to save everyone time, so use it as a starting point for questions about what each stage assesses and how that capability shows up in the job.
Approach the conversation a little like an investor: what does this company prioritize, how does it make decisions, and do its practices reflect what it says it values?
Could we pass our own evaluations?
If you’re hiring engineers at your organization, try a small experiment: give your current team the assessment you use for new candidates. Give them the same amount of time, the same tools and limitations, and the same information. Would everyone (or anyone) pass?
If someone we trust to deliver important work struggles with our own hiring assessment, what does that tell us? Is there a capability they should strengthen? Or is the assessment giving too much weight to something that has little bearing on their real contribution?
The results wouldn’t automatically prove the test is good or bad. But they would give us a reason to compare what we’re measuring with what we already know about someone’s work, rather than keeping a format simply because it’s familiar or provides a score.
Being a candidate again has made me much more interested in the assumptions behind the processes I’ve used myself.
The "Real Life Case Study"
At my current org, I favor a technical assessment built around a realistic case study or scenario. It gives candidates room to think, ask questions, and discuss situations they might actually encounter in the organization. You can explore what they need to know before making a recommendation, and whether they can explain the consequences of a technical choice. It feels more like a conversation than an interrogation.
The choice of problem matters, too. A generic exercise, such as “design a notification system" gives everyone a common starting point, but can easily be rehearsed. A problem relevant to the business lets us explore how someone connects technical decisions to customer needs, existing systems, and the capabilities of the team.
We need to provide enough context to make that fair. Candidates shouldn’t have to guess how our business works. And while several approaches may be reasonable, we still need clear evaluation criteria to distinguish sound reasoning and awareness of risks from a confident delivery.
The Most Valuable Quality: A Learning Mentality
One of the most valuable qualities I look for when hiring in the AI era is a learning mentality. I want people who are curious about what they don’t yet understand and willing to examine their assumptions.
We can observe this during the interview. Introduce a new constraint into the case study, or question an assumption behind their recommendation. Maybe the team is smaller than they expected, or a dependency can’t provide the guarantee their design relies on. How do they respond?
I wouldn’t expect someone to abandon their position just because I challenged it. They might have a good reason to defend it. What interests me is whether they explore the new information and explain whether it changes their thinking. For me, that willingness to learn from someone else is a practical sign of humility.
Being able to say “with that context, I’d approach it differently” is a useful signal. So is explaining why the original approach still holds up. Those are behaviors I want to see when a team encounters something unfamiliar or disagrees about how to move forward.
I have a hard time separating those things into “technical skills” and “soft skills.” Explaining why a design creates unnecessary risk requires both. So does disagreeing with a proposed solution without shutting down the conversation.
In my previous article, I wrote about AI helping people contribute beyond the boundaries of their usual roles. If we expect engineers to take broader responsibility for products and business outcomes, we should make room to assess that broader contribution.
So....would I hire me today???
I hope so, but at a minimum, I’d want a solid process that gave me enough evidence to answer honestly. And before asking the next candidate to go through it, I’d want to know whether it would recognize the capabilities and skills I value in the people who are already doing the job well.
share the love
The views and opinions expressed on this blog are the author's own and do not reflect those of their employer, past or present. Any content shared here is for informational purposes only and should not be taken as professional or legal advice.
The Afterburner
The conversation continues with additional thoughts, insights and commentary.One thing I didn't touch on in the original article was the use of psychometric assessments as an early hiring gate. I recently encountered Bryq, which evaluates cognitive abilities and personality traits before candidates progress through the interview process. I passed the assessment, but it left me questioning whether this is really the best way to identify talent.
I can appreciate the value of understanding how someone thinks and approaches problems. But what about candidates who are excellent at their jobs yet struggle with abstract logic tests under time pressure? Or neurodivergent candidates whose strengths might not translate well into this particular format? When these assessments become the first hurdle, we risk filtering out capable people before we've even had a conversation with them.
There's also the question of whether we're unintentionally favoring candidates who fit a particular cognitive or personality profile, rather than recognizing different ways of being effective. I'm not against psychometric assessments, but I question their place as a gatekeeper. They might provide useful additional context, but I'd much rather use them to inform a hiring conversation than determine whether someone deserves one in the first place.