With IT firms and employees increasingly using AI tools to write codes, hiring tests for entry level and junior engineers are getting changed. From being asked to write simple codes, candidates are being asked to optimise and restructure existing AI written codes, to improve quality, say analysts working on AI and cyber security. With AI tools easily available, IT firms are also ensuring that they don’t hire candidates who can just copy paste codes.

Experts say traditional tests that require candidates to write code from scratch don’t reflect the realities of modern development, especially in large organizations with extensive code libraries and AI tools.

Chirag Mehta, vice president and principal analyst, Constellation Research, said, “The focus in hiring has shifted from just programming skills to broader problem-solving capabilities….. Hiring tests for entry-level engineers are evolving to gauge candidates’ ability to learn new methods, optimise existing code, and “refactor” it to improve quality. It’s about finding engineers who are agile thinkers, not just scriptwriters.” Refactoring is restructure the source code of an application or piece of software so as to improve operation without altering functionality.

Mehta added that with Generative AI automating a big chunk of programming tasks, companies aren’t just looking for programmers anymore. They need engineers who can innovate and adapt. They are looking for engineers who fall in love with problems and not solutions.

On questions about the need to secure AI-generated code, Mehta elaborated that AI-generated code introduces a unique security challenge. “There’s a bias to trust code created by AI, but these tools are as prone, if not more, to introducing vulnerabilities as human coders. If an AI tool generates insecure code, that risk multiplies as more developers rely on it across projects. Companies must “shift left” to address security upfront—reviewing AI-generated code for vulnerabilities and rigorously validating the tools themselves. Security needs to be baked into the code generation process, ensuring that AI aids development without compromising safety, leading to security by design.”

Yugal Joshi, partner, Everest group, said, “Unlike earlier times where developers were focused more on technical aspect of software development, now other skills such as reviewing AI-generated code, assessing AI-generated documents such as, code comments, requirements, etc is important as well.” He also added that a lot of these depends on the seniority for which the software developers are being hired.

The change in the pattern of tests for software developers in IT firms is coming at a time when companies like Infosys, Cognizant, Google, SAP and others have recently boasted about their AI written codes, even as other IT firms like TCS spoke about having large libraries of codes for their employees to make use of. Gaurav Vasu, founder of tech market intelligence firm UnearthInsight, said, by 2025-2026, about 10-15% of new code will be written by AI / Gen AI in the IT services industry as AI models are getting commoditized and cost-effective.

Vasu added, “Freshers exams amongst top IT firms have recently seen increased weightage being given (about 20-30%) to logical reasoning, domain and deep Tech questions. This is happening because we are seeing increased momentum in software automations. With AI tools easily available, IT firms are ensuring that they don’t hire candidates who can just copy paste codes.”

He also added that calendar 2025 onwards test would start including questions on knowledge of AI Coding tools as they will become coding assistants for freshers in future.

Nandini Tandon, cofounder and chief people officer, Indusface, a cyber security firm, said, that in field of software development and cyber security that is increasingly being influenced by AI, freshers must understand not only code but also the ethical implications and vulnerabilities introduced by AI tools. “We look for those who can adapt to rapid advancements, think critically, and stay vigilant in identifying AI-induced risks. This shift highlights a need for well-rounded talent equipped to protect evolving digital landscapes, going beyond traditional coding knowledge.”

With IT firms and employees increasingly using AI tools to write codes, hiring tests for entry level and junior engineers are getting changed. From being asked to write simple codes, candidates are being asked to optimise and restructure existing AI written codes, to improve quality, say analysts working on AI and cyber security. With AI tools easily available, IT firms are also ensuring that they don’t hire candidates who can just copy paste codes.

Experts say traditional tests that require candidates to write code from scratch don’t reflect the realities of modern development, especially in large organizations with extensive code libraries and AI tools.

Chirag Mehta, vice president and principal analyst, Constellation Research, said, “The focus in hiring has shifted from just programming skills to broader problem-solving capabilities….. Hiring tests for entry-level engineers are evolving to gauge candidates’ ability to learn new methods, optimise existing code, and “refactor” it to improve quality. It’s about finding engineers who are agile thinkers, not just scriptwriters.” Refactoring is restructure the source code of an application or piece of software so as to improve operation without altering functionality.

Mehta added that with Generative AI automating a big chunk of programming tasks, companies aren’t just looking for programmers anymore. They need engineers who can innovate and adapt. They are looking for engineers who fall in love with problems and not solutions.

On questions about the need to secure AI-generated code, Mehta elaborated that AI-generated code introduces a unique security challenge. “There’s a bias to trust code created by AI, but these tools are as prone, if not more, to introducing vulnerabilities as human coders. If an AI tool generates insecure code, that risk multiplies as more developers rely on it across projects. Companies must “shift left” to address security upfront—reviewing AI-generated code for vulnerabilities and rigorously validating the tools themselves. Security needs to be baked into the code generation process, ensuring that AI aids development without compromising safety, leading to security by design.”

Yugal Joshi, partner, Everest group, said, “Unlike earlier times where developers were focused more on technical aspect of software development, now other skills such as reviewing AI-generated code, assessing AI-generated documents such as, code comments, requirements, etc is important as well.” He also added that a lot of these depends on the seniority for which the software developers are being hired.

The change in the pattern of tests for software developers in IT firms is coming at a time when companies like Infosys, Cognizant, Google, SAP and others have recently boasted about their AI written codes, even as other IT firms like TCS spoke about having large libraries of codes for their employees to make use of. Gaurav Vasu, founder of tech market intelligence firm UnearthInsight, said, by 2025-2026, about 10-15% of new code will be written by AI / Gen AI in the IT services industry as AI models are getting commoditized and cost-effective.

Vasu added, “Freshers exams amongst top IT firms have recently seen increased weightage being given (about 20-30%) to logical reasoning, domain and deep Tech questions. This is happening because we are seeing increased momentum in software automations. With AI tools easily available, IT firms are ensuring that they don’t hire candidates who can just copy paste codes.”

He also added that calendar 2025 onwards test would start including questions on knowledge of AI Coding tools as they will become coding assistants for freshers in future.

Nandini Tandon, cofounder and chief people officer, Indusface, a cyber security firm, said, that in field of software development and cyber security that is increasingly being influenced by AI, freshers must understand not only code but also the ethical implications and vulnerabilities introduced by AI tools. “We look for those who can adapt to rapid advancements, think critically, and stay vigilant in identifying AI-induced risks. This shift highlights a need for well-rounded talent equipped to protect evolving digital landscapes, going beyond traditional coding knowledge.”