The Goodwin Blog

Stay Up To Date With The Latest News & Insights

I’ve spent most of my career helping companies hire software engineers. It started with grad school in computer science, data science, and informatics, and today I’m a Recruiting Partner at Goodwin Recruiting, focused on software engineers and broader IT talent.
So here’s my question for hiring managers: should you include coding assessments in your interview process?
Yes. With one big caveat: don’t block candidates from using outside resources.

Why locked-down assessments backfire

Most candidates run into a coding test at some point. What surprises me is how many are expected to code from scratch, with no browser tabs, no AI tools, and no access to certain sites. Some call these “blocked” assessments.
Here’s the thing. Nobody codes from scratch in real life. Everyone uses outside resources. You need engineers who are resourceful and can perform in the real world, and a blocked test doesn’t show you that.
AI makes this even more important. You’re not just testing whether someone can write code anymore. You’re testing whether they can use AI tools well, and whether they actually understand what the AI produced so they can catch it when it’s wrong.

More reasons to skip the blocked test 

  • It doesn’t show real skill. Nobody memorizes code anymore. What matters is whether a candidate can build the solutions your team needs.   
  • You’ll lose good candidates. Plenty of engineers will walk away the moment they hear a blocked assessment is part of the process. They could be exactly who you want, and you’ll never know they were interested.
  • The data can mislead you. Some people crush restricted tests and struggle on the job. Others bomb them and would have been great hires.

You’re not hiring someone to memorize code. You want someone who sees the big picture and solves real problems with their experience and the tools available.

What assessments platforms can (and can’t) tell you 

Codility, HackerRank, CodeSubmit, and HackerEarth are all popular, and some can be set up as locked environments. They’re fine for baseline proficiency. But they can’t tell you how deep a candidate’s experience goes, how they collaborate, or how they make engineering decisions. For that, you need a real conversation.

How I would run a better technical interview

  • Talk through their experience. Have them walk you through a big project: their role, the tech they used, the challenges, and the results.
  • Give them a realistic problem. Something close to what they’d actually face at your company. You’ll get a much better read on their coding, their judgment, and how they deliver practical solutions.
  • Throw out a hypothetical. A system failure, an architecture disagreement, a performance issue. Ask how they’d approach it, what they’d prioritize, and who they’d pull in. You’ll see how they think, how they communicate, and how they balance tech with business needs.

As my colleague David Ivey, Salesforce Administrator at Goodwin Recruiting, puts it: “It’s always best to talk things through. It’s okay to use AI tools, as long as the candidate proves they deeply understand the overall process. Proving you can code matters, but since AI can help with that now, you need someone who can see AI’s blindspots and tell whether the AI’s answer is right for your business.”

Let us help you find the right engineer 

When you work with the Goodwin Recruiting IT Team, you get fully vetted candidates who fit your needs and your culture. We dig into how candidates think, apply their experience, collaborate, and make technical decisions, because those are better predictors of success on the job than whether they can write code under pressure.

Ready to meet strong software engineers or other specialized IT pros? Reach out. I’m Scott Strong, and I’d love to help you find your next hire.