Episode 225
Technical Assessment in the Age of AI: Rethinking Hiring
Whiteboard interviews are dead. As AI reshapes how engineers work, companies must rethink what they’re actually assessing for—and whether their interview process reflects the job. Amanda Richardson, CEO of CoderPad, breaks down the evolution of technical hiring.
Episode Key Takeaways
The whiteboard interview persists as habit, not science. Researchers found that adding an audience to a whiteboard problem cut pass rates by two-thirds, with women disproportionately affected—a stark reminder that assessment design shapes who gets hired, not just how well they code.
False negatives cost more than false positives. Focusing solely on weeding out weak candidates means overlooking talent that competitors will hire. CoderPad’s first developer had a biochemistry degree and no CS background—she would have been rejected on resume alone but became a core product builder.
AI isn’t cheating; it’s a skill to assess. Companies banning AI in interviews while requiring AI fluency on the job create a contradiction. The better approach: be explicit about when AI is allowed, track its use, and evaluate how candidates prompt, validate, and think critically about AI outputs.
Systems design and communication now matter more than syntax. As AI handles routine coding, hiring teams should assess how candidates break down problems, explain trade-offs, and collaborate—skills that actually predict job performance, not memorization of algorithms.
Disconnect between hiring and L&D definitions of AI skills creates misalignment. Many organizations lack a shared rubric for what ‘strong AI skills’ means across recruiting and learning teams, leading to hiring for undefined competencies and onboarding friction.
Frequently
Asked
Questions
Is it cheating to use AI during a technical interview?
It depends on clarity and policy. If a company requires AI fluency on the job but bans it in interviews, that’s contradictory. The solution: explicitly state when AI is allowed, record its use, and assess whether candidates prompt effectively, validate outputs, and think critically—not whether they avoid it.
What skills should replace whiteboard coding in technical interviews?
Systems design, communication, and trade-off reasoning. Rather than syntax recall, evaluate how candidates break complex problems into components, explain their choices, consider edge cases, and collaborate. These predict real job performance and reflect how engineers actually work with modern tooling.
How do I avoid false negatives in technical hiring?
Use work samples and pair programming over resume screening. A two-hour collaborative project reveals skills that CVs and logos hide. Companies often reject candidates who later build successful products elsewhere. Asynchronous projects with follow-up conversations replicate real work better than watched, high-pressure interviews.
What's the candidate experience impact of a poorly designed assessment?
If 60% of candidates fail your top-of-funnel assessment, ask whether the product is hard to use or the skill bar is genuinely high. Aim for 95–100% completion rates, then calibrate scoring to role requirements. Poor UX screens out good candidates for the wrong reason.
How do I align hiring and L&D on AI skills definitions?
Start by asking hiring managers to define ‘strong AI skills’ using performance reviews of current employees. Connect with L&D on their upskilling rubric. Without shared definitions, you hire for undefined competencies and create onboarding misalignment. Explicit alignment prevents waste.