AI tools widen gap in junior software engineers' skills

BairesDev has published a global survey on how junior software engineers use artificial intelligence tools. The findings point to a gap between how junior developers judge their own readiness and how senior engineers assess it.

The survey covered 1,569 developers across 77 countries, including 1,059 junior engineers and 510 senior engineers. It found that 85% of junior developers say AI tools have improved their understanding of software development, while only 16% of senior developers believe juniors fully understand the AI-generated code they submit.

The results add to a wider debate in the software industry over whether AI coding tools are helping new entrants learn faster or weakening core technical skills. One concern is that developers may be able to generate working output without being able to explain it, debug it, or rewrite it independently.

That tension appeared clearly in responses on confidence and hiring priorities. Nearly a quarter of junior developers said writing code from scratch was the task they felt least confident performing without AI assistance, and only 5% said that ability was critical to getting hired today.

Instead, 48% of junior respondents ranked problem-solving and analytical thinking as the top skill for getting hired, compared with 18% who cited AI tool proficiency. Among senior developers, 72% identified critical thinking as a foundational skill for junior developers over the next three years.

Readiness gap

The survey suggests the disagreement is not over whether AI matters in software work, but over what should count as job readiness. Senior developers placed greater weight on practical experience and applied judgment than on fluency with AI tools.

Seventy per cent of senior developers ranked real-world project experience as the strongest indicator that a junior developer is ready to contribute. That was followed by internships at 56% and strong performance in practical coding tasks at 53%.

Junior developers broadly agreed on the value of practical exposure. Half said education should provide more real-world project experience, making it the most common response by a wide margin.

The data also showed mixed views among experienced engineers on graduate readiness more generally. Just over half, 51%, said graduates are broadly ready to contribute after onboarding. Another 38% said there is a clear gap between university training and real-world application, while 11% said graduates need significant additional training before they can contribute.

Nacho De Marco, Chief Executive Officer and Co-Founder of BairesDev, said the role of coding itself has shifted as AI tools become more common in development work. "For most of software engineering's history, writing code was the job. It was the craft, the proof of skill, the language developers used to think. AI has changed that. Generating code is no longer the hard part - and that means it can no longer be the measure," De Marco said.

He said the main issue was the distance between output and understanding. "What this data is telling us is that the next generation of developers is learning to produce output without fully owning it. That gap between generating and understanding is exactly where the profession needs to focus. If we don't close it now, we have a real problem: where are the senior engineers, architects, and technical leaders of 2030 and 2035 going to come from? The seniors of the future are the juniors of today. And right now, the people closest to the work are telling us the foundation isn't there yet," he said.

Skills debate

An academic review of the findings reached a similar conclusion on the importance of enduring technical judgment. The analysis was written by Professor Francisco Anello, Director of the Master in Business and Technology at Universidad de San Andrés.

"The most revealing finding in this study is not the gap between junior and senior developers - it is where they agree," Anello said.

"Critical thinking and deep code comprehension rank far above AI tool proficiency. These are the same competencies that technical training has always prioritized," he said.

The findings arrive as employers, universities, and training providers adjust to AI becoming part of routine software work. The survey points to an industry that may now place less emphasis on writing every line of code manually and more on understanding, testing, and taking responsibility for machine-assisted output.

It also indicates that junior developers do not see AI tool use as the main route into employment. Their responses suggest they still place the highest value on analytical ability, even as they rely heavily on AI in day-to-day coding tasks.

BairesDev said the Dev Barometer has collected responses from more than 4,000 engineers across more than 70 countries since 2025. In this edition, the contrast between confidence in AI use and concern over basic coding independence stood out most sharply, with 24% of junior developers saying they are not confident writing code from scratch without AI assistance.