On Naukri and LinkedIn right now, a strange split shows up in the data: senior hiring is holding steady, even growing, while the entry-level door that Indian graduates have walked through for two decades is quietly narrowing. Staffing firm Xpheno puts it in blunt numbers — net hiring in India’s IT sector fell from roughly 6 lakh positions in FY22 to about 1.4 lakh in FY26. Nothing about that is a prediction. It’s already happened. The question for a student still in college isn’t whether this shift is real, but what to actually do about it, because “learn AI” has become the least useful advice anyone can give.
The entry-level squeeze, not the robot apocalypse
The mistake is picturing AI replacing whole jobs. What’s actually happening in Indian IT is narrower and, in some ways, more unsettling: AI is eating the specific tasks that used to justify hiring a large batch of freshers in the first place. A Nomura-based report found about 55% of Indian IT companies had cut entry-level hiring, against just 14% for senior roles — because software testing, basic coding, and documentation, the traditional on-ramp for a junior employee, are exactly what current AI tools are good at. NASSCOM’s own numbers show tech-sector workforce growth slowing to 2.3% in FY26, even as company revenues kept climbing. Companies aren’t shrinking. They’re hiring differently.
What the global data says is coming by 2030
Zoom out and the picture is less bleak, though it demands more from individual students. The World Economic Forum’s Future of Jobs Report 2025, based on over 1,000 employers across 55 economies, estimates that 92 million jobs will be displaced by 2030 but 170 million new ones created — a net gain of 78 million roles globally. The catch is what it calls a skills timeline: nearly 40% of the skills workers currently use are expected to change or become outdated in that same window. The report’s “core skills for 2030” list leans less on any single tool and more on a mix — analytical and creative thinking, technology literacy, and resilience and adaptability all rank near the top. Automation doesn’t remove the need for skill. It just shortens how long any one skill stays useful.
Fluency beats a certificate
Where this gets concrete for an Indian engineering student is employability data that predates the AI conversation but now compounds it. The Wheebox ETS India Skills Report found only about 42.6% of Indian graduates meet basic industry employability standards — a gap that was already there before AI, now sharpened by it. What’s changed is what closes that gap fastest. Hiring managers increasingly care less about whether a candidate has heard of GitHub Copilot or ChatGPT and more about whether they’ve used such tools inside a real project, debugged the parts the AI got wrong, and can explain why a particular approach was chosen. That last part — explaining and defending a decision — is hard to fake and hard to automate, which is precisely why it’s becoming the filter.
Build proof, not just knowledge
None of this points toward abandoning a computer science degree or chasing every new AI buzzword. It points toward treating college years as a window to build a small, honest portfolio: two or three real projects, built with AI tools rather than despite them, that a student can open in an interview and walk through line by line. Domain knowledge still matters — a data analyst who understands the business problem, not just the SQL query, is the one AI struggles to replace. So does the unglamorous combination of communication, adaptability, and the willingness to keep re-skilling, because the WEF’s 40% figure isn’t a one-time hurdle to clear before 2030; it’s the new baseline pace of change. The freshers getting hired in 2026 aren’t the ones who avoided AI or the ones who only talk about it. They’re the ones who can show what they built with it.