The landscape of software engineering has experienced a seismic shift due to artificial intelligence, making traditional coding assessments increasingly outdated. As a result, some organizations have banned the use of AI tools during evaluations to mitigate cheating concerns.
The function of a software engineer is evolving, emphasizing strategic decision-making rather than mere code writing. Aspiring engineers are finding the job market tougher than ever, with significant job losses in the industry compounding the competition for available positions. The surge of AI technology has triggered worries about interview integrity, prompting companies to adapt their priorities almost daily in response to rapid advancements.
A pressing challenge for hiring managers lies in assessing the suitability of candidates now that AI can generate code autonomously. Career experts and software engineers indicate that conventional interview methods are out of sync with how AI has transformed the daily roles of programmers, complicating the hiring process for both applicants and employers.
Stefan Mai, a former engineer at Meta and Amazon, characterized AI’s impact on engineering interviews as akin to an “atomic bomb.” Software engineering is among the first sectors where AI’s influence is palpable. A report from Google's research division revealed that 90% of tech professionals utilized AI for tasks such as writing and refining code, marking a 14% increase from the prior year. This adaptation has made the software engineering field a key indicator of broader AI integration within industries.
AI’s Role in Shaping Software Engineering Professions
Today, AI assists software engineers in writing code, creating documentation, analyzing data, learning programming concepts, and troubleshooting various challenges, thereby enabling tech firms to accelerate their workflows. For instance, an engineer at OpenAI recently expedited a system overhaul that would have taken his team an entire week to execute, a shift highlighted by company president Greg Brockman in a discussion with Sequoia Capital. Furthermore, many internal applications at Google are reportedly developed predominantly using the company’s Antigravity AI coding tool, according to Varun Mohan, a director at Google DeepMind.
Boris Cherny, who leads Claude Code at Anthropic, shared on X that his output for the product was entirely generated by Claude Code over the last month. He posits that AI is redefining the software engineer’s role, steering it towards higher-level strategic tasks instead of just coding. He suggests that the term “software engineer” could soon be supplanted by a title like “builder” to accurately reflect the evolving nature of the work.
Google’s Mohan clarified that AI is not intended to replace human engineers, emphasizing the need for developers to focus on the critical question of what should be built. Madhu Kurup, Indeed's vice president of engineering, compared AI’s role in software engineering to that of Google Maps in navigation—providing guidance without determining routes or destinations.
Despite this technological advance, the Google report noted that 46% of tech workers are only somewhat confident in the quality of AI-generated code, while 31% feel AI has only marginally improved coding practices. Layoffs projected through 2025 and 2026 reflect AI's impact on employment trends, with the technology cited as a primary reason for job cuts by Challenger, Gray & Christmas earlier this month.
A Shifting Assessment Landscape
Traditionally, candidates have been evaluated through rigid testing protocols akin to standardized exams for coding. However, these methods fail to capture how modern engineers leverage AI tools to delegate tasks, navigate problems, and enhance efficiency. As a result, some engineers argue that these tests no longer accurately depict the responsibilities they will encounter in actual jobs.
Software developer David Barajas noted that during recent interviews, none of the companies inquired about his use of AI coding tools like Cursor in his work. Instead, he remarked that interviewers often prohibit any AI assistance during assessments, even requiring him to share his screen to validate compliance.
Sujata Sridharan, a seasoned software engineer from fintech firm Bolt, echoed this sentiment, pointing out that many firms continue to employ outdated testing methods that prioritize understanding code over AI collaboration. She highlighted a growing disconnect between the skills tested and the realities that engineers face in their roles.
Concerns regarding cheating in the age of AI have intensified, with barriers against using AI tools during tests becoming more pronounced. This rapid evolution in technology means that role specifications are ever-changing; for instance, a company hiring for Ruby on Rails might shift priorities within weeks as AI makes it easier to translate code from other languages.
New Strategies for an Evolving Challenge
In response to these developments, employers are increasingly prioritizing questions that reveal candidates' thought processes and decision-making skills rather than pure coding abilities, a trend typically seen in senior interviews that is now widespread. Some innovative startups have even invited candidates to partake in half-day onsite evaluations. Moreover, it is becoming more common for employers to allow the use of AI in assessment scenarios.
Despite these adjustments, many hiring practices still fall short in encapsulating the modern workflow. Sridharan indicated that her typical collaboration with AI focuses on problem-solving, yet interview settings often reduce it to mere coding tasks. Acknowledging the complexity of this issue, Mai described the current state of candidate assessment as an “unsolved problem,” highlighting the unpredictability of challenges candidates face in interviews.
