What Do Tech Companies Want From New Hires?

Insights from CodePath.org 2025 Early-Career Hiring Survey 

How is the rise of AI affecting hiring decisions in the tech industry? How much do reputation and pedigree matter in hiring decisions? What skills—technical or otherwise—are most important for success on the job? 

CodePath asked engineering leaders actively engaged in hiring across a variety of industries to identify the most critical signals, skills, and challenges related to hiring early-career engineers.

Methodology

From August to September 2025, CodePath surveyed 222 engineering leaders actively engaged in hiring across a variety of industries.

Respondents represented a range of industry categories and employer sizes (<10-5000+ employees), with a quarter of respondents from companies with over 5,000 employees. Respondents were asked a mix of multiple-choice and open-ended questions. Data are reported as rounded percentages based on the number of respondents who answered each question. For questions allowing multiple selections, totals may exceed 100%.

Key Findings

Survey findings paint a picture of rapid, AI-driven transformation where non-technical skills and work experience often surpass the value assigned to technical skills or academic pedigree in the hiring process.

• Entry-level hiring for engineering roles has slowed, not stalled.
  • While 38% said they have reduced the number of entry-level hires over the past year, only 13% of respondents had paused all entry-level hiring. In fact, 8% report increasing early-career talent. One in five (18%) report no change. 
    
    
• AI skills are a prerequisite for hiring.
  • Half of respondents report asking early-career candidates about their use of AI tools as part of the interview process. 
  • 55% of respondents say they now expect candidates to demonstrate greater fluency with AI tools and frameworks.  
  • Anecdotes reflect shifts in hiring practices.
    • As one respondent put it, they are “transitioning from a policy of ‘using AI in the interview is not allowed’ to ‘using AI in the interview is required.’”
    • Another said, “We hired 1,000 interns last year. We will do it again! Early-career are ‘AI-native’ and we want that.” 

• AI is upping expectations for early talent hiring. 
  • About half of leaders say expectations have risen over the past two years, including 21% who cite AI specifically as the reason for raising the bar on skills expectations. Only 22% of respondents say their expectations are about the same, and no one reported lowering their expectations. 
  • When asked about changing skills expectations of early-career hires, greater fluency with AI tools and frameworks was the most common response (55%), followed by faster time to writing production-ready code (39%), ability to learn new tools or languages more quickly (34%), improved debugging independence (33%) and familiarity with modern development workflows (28%).
  • One respondent said, “The bar for entry level changed, really changed. We expect entry level to have similar skills as someone with 2+ years and also real hands-on experience building with AI native tools. Yes, it’s unfair, but that’s the reality.” 

• The ability to evaluate AI output is key.

• Fifty-six percent of employers said critically evaluating AI outputs is very or extremely important, while using AI for debugging was the second highest ranked skill, with 48% of respondents saying it is very or extremely important.
• Open-ended responses indicated that “understanding when not to rely on AI,” knowing AI’s limits, and “being able to justify and explain code changes, whether you created the code with or without AI help” are all important skills employers expect among early-career talent.

• Technical skills gaps persist, with debugging and maintainable code topping the list of employer concerns.

• Over half of respondents report skill gaps in writing clean, maintainable code (57%) and a similar number note gaps in debugging and troubleshooting (53%). 
• These two skills were also the top rated skills sought by employers, with two-thirds of employers indicating they are very or extremely important.

• Demonstrated skills matter more than college pedigree.
  • The top readiness signals for employers outside of the interview process are side projects or portfolios (38%), internship experience (35%), and public code portfolios like GitHub (34%).
  • In contrast, credentialing programs (4%), type of degree or academic focus (23%), and school prestige (17%) carry far less weight in hiring decisions.

• Soft skills are still the best predictors of success.
  • Employers highlight ownership, initiative, and communication as the most critical non-technical traits for early-career engineers to thrive. 
  • While technical interviews were the most common signal for assessing candidate readiness for a role (54% of respondents put it in their top three), behavioral interview responses were a close second, at 48%. Personal or side projects came in third, at 38%. 
  • Employers are eager for workers who can communicate, work across other teams (e.g., product or design) and who are curious and willing to take on new tasks, yet a majority of employers said their new hires lack confidence or fear failure (55%) or have difficulty breaking down and tackling new, large tasks (54%), and 42% said new hires are hesitant to ask for help or clarification.

• While hybrid and remote workplaces remain the norm, employers see value in being in person at least some of the time.

• Thirty-two percent of respondents said physical proximity to the office is “very important,” though only 9% of companies in the survey were fully in person.   
• A majority of those working in hybrid or remote settings say lack of proximity limits informal learning opportunities like pair programming (53%), and nearly half say it creates challenges in building relationships with peers or mentors (48%).
• Engineering leaders had this to say: 
  • “Biggest issue is building peer networks without in-person time.”
  • “Early-career hires often feel disconnected when remote.”
  • “You don’t overhear or pick things up as easily when remote.”