CoderPad: U.S. Technical Hiring Is Up 90% in 2026. The Interview Format Has Changed.
New Trends in Tech Hiring
Tech hiring feels contradictory right now. Layoffs keep making headlines, including more than 45,000 global tech job cuts in Q1 2026 alone. At the same time, CompTIA’s State of the Tech Workforce 2026 report, published this month, projects 185,499 net new tech jobs in 2026 after a 0.3% dip in 2025. Both things are true, and the tension between them is creating a confusing environment for candidates.
What’s less discussed is what the recovery looks like from inside the hiring process itself. A March 2026 survey of 650+ developers, recruiters, and hiring leaders published by CoderPad offers the clearest picture yet. Technical assessments are up 48% globally compared to mid-2023. In the U.S., technical hiring activity has increased 90%.
The volume is there. But the interview that candidates are preparing for has quietly been replaced by something different.
Hiring Is Rebounding, and the Bar Has Been Raised
The CoderPad data cuts against the narrative that AI is suppressing technical hiring. Companies aren’t pausing on engineers. If anything, they’re investing more in finding the right ones: 60% of talent leaders say improving quality of hire is their top priority this year, and 53% expect their hiring budgets to increase, the highest share in years.
What’s changed is what “quality” means. Companies are no longer hiring for roles centered on writing boilerplate, handling low-complexity implementations, or solving problems that AI tools can already draft a reasonable answer for. They’re hiring for roles where human judgment, architecture thinking, and the ability to work with AI as a collaborator are the actual job. The interview has started to reflect that.
This is the context behind a significant structural shift: 92% of companies reported an increase in hiring for AI-related positions in early 2026, according to data cited by TechTimes. Those roles carry a 56% wage premium, according to CompTIA. Companies aren’t just hiring more; they’re willing to pay more for a specific capability profile.
Meta’s AI-Enabled Coding Round Is the New Template
The clearest example of what updated hiring looks like in practice comes from Meta. In October 2025, Meta replaced one of its traditional algorithmic interview rounds with a new AI-assisted coding round using CoderPad.
The format: a 60-minute session where candidates work with a real codebase and have access to AI assistants, including models like Claude Sonnet, Gemini 2.5 Pro, and GPT-4o mini. The tasks aren’t algorithmic puzzles with clean inputs and expected outputs. They involve building features on top of partially implemented code, debugging broken implementations, and explaining design trade-offs and complexity after the work is done.
The critical difference from a LeetCode prep loop is that the interview isn’t testing whether you can recall a solution pattern. It’s testing whether you can direct AI tools effectively, catch errors in generated code, explain the reasoning behind your choices, and operate like someone doing real engineering work. Preparing for this format requires a different kind of practice than grinding easy-to-hard problems in isolation.
What Companies Are Now Testing
The CoderPad report identifies the specific scenarios that hiring teams are moving toward. The list is concrete:
- Debugging AI-generated code with identifiable errors or logic gaps
- Explaining system design decisions and discussing trade-offs between approaches
- Iterating on AI output rather than producing code from scratch
- Communicating what the code does and where its limits are
These tasks don’t reward memorization. They reward clear thinking under pressure, familiarity with how AI tools behave and fail, and the ability to explain work to another person.
The 82% of developers who told CoderPad that GenAI is useful in their work, and the 54% who said their productivity would drop at least 10% without AI tools, aren’t outliers. They represent the baseline of how modern engineering work operates. Companies are building interviews that reflect that reality. The IQ AI Interviewer is built for exactly this kind of adaptive practice, including debugging and problem decomposition scenarios.
What This Means for Candidates Preparing Now
The shift in format changes what good preparation looks like. A few practical implications:
Debugging matters as much as building. If your prep is all about writing solutions to new problems, you’re training half the skill set. The ability to read unfamiliar code, identify the issue, and explain the fix is being tested directly.
Trade-off conversations have moved earlier. System design thinking used to be primarily a senior-level concern. The CoderPad data suggests it’s now showing up in how teams evaluate mid-level candidates as well, through scenario prompts and design discussion components.
AI tool fluency is expected but not sufficient. Companies aren’t grading whether you used AI; they’re grading whether you validated and directed it well. Knowing when AI output is wrong is as important as using it.
The 28% of hiring teams that CoderPad reports are actively growing their early-career pipelines are doing so with these formats. Candidates entering the market now are being assessed against a higher bar of AI-aware judgment, not a lower one. Working with a coach to run through these scenarios before your loop can close the gap quickly.
The Bottom Line
The CoderPad State of Tech Hiring 2026 data and Meta’s October format change point in the same direction. Technical hiring volumes are recovering, but the interview has been redesigned for the way engineering work actually happens now.
The old technical interview tested memory, speed, and pattern recognition. The new one tests judgment, reasoning, and the ability to work with AI as a tool rather than a crutch. These are different skills, and they require different preparation.
Candidates who spend the next few months grinding LeetCode for the ghost of an interview that no longer exists will find themselves underprepared. Those who practice in formats that mirror what companies are now running, including AI-augmented coding sessions and system design discussions, won’t.