For decades, the first year of an engineering career came with enough room to be inexperienced. Companies gave young recruits manageable problems and senior colleagues to turn to. In return, they built a pipeline of engineers who understood their systems well enough to take responsibility for them. Training a beginner was an investment in the people the business would eventually need.
Artificial Intelligence (AI) is taking over much of that work. For a company, the immediate gain is clear. Routine tasks get done faster. Which is why companies are not abandoning the entry-level role, but they are no longer willing to hire someone and spend a year teaching them. For a young engineer, the loss runs deeper. Each task automated is also a chance to learn how a system behaves, where it breaks and what a careless decision can cost. This leaves the industry with a question that reaches well beyond this year’s hiring targets: How will beginners acquire the experience employers increasingly expect them to arrive with?
The numbers point the same way. An EY analysis from 2025 estimated that entry-level IT roles in India had declined by 20–25% as companies automated and hired more narrowly. PwC’s 2026 AI Jobs Barometer, which looked at more than one billion job advertisements, found that the junior roles most exposed to AI are seven times more likely to ask for skills once associated with senior staff, such as leadership and strategic thinking. Early-career postings in highly exposed sectors have flatlined, while entry-level roles that demand experienced-level skills have grown 35% since 2019.
The first job is becoming a test of abilities it once helped people develop. Engineering education has to reckon with that shift. I see this in our own hiring at Classplus. The junior engineers who stand out are the ones who question what AI hands them. They read generated code critically, ask why a system behaves the way it does, and have a view on what is worth building. In interviews, we now spend less time on whether someone can write a function and more on whether they can explain, test and defend one.
{{/usCountry}}The first job is becoming a test of abilities it once helped people develop. Engineering education has to reckon with that shift. I see this in our own hiring at Classplus. The junior engineers who stand out are the ones who question what AI hands them. They read generated code critically, ask why a system behaves the way it does, and have a view on what is worth building. In interviews, we now spend less time on whether someone can write a function and more on whether they can explain, test and defend one.
{{/usCountry}}Engineering teams elsewhere describe the same squeeze. Reporting in The New Stack points to AI-generated code creating review bottlenecks, so more of a team’s day goes to checking and deciding than to typing. A Linux Foundation report cited there found a 27% net rise in tech hiring across Europe, yet entry-level technical hiring fell 3%. Companies appear to be hiring fewer juniors and more mid and senior people who can exercise judgement and oversee AI systems.
So the junior role is being judged on what used to be a senior engineer’s yardstick, at the exact moment the on-the-job training that used to get people there is disappearing. That raises a longer-term question: Who will develop the experienced engineers everyone wants to hire?
Engineering education now carries a greater share of that responsibility. Students need opportunities to make consequential decisions while they still have the time and guidance to learn from them.
India’s engineering education has real catching up to do. A 2026 baseline study of B.Tech Computer Science and allied curricula, conducted by MeitY with industry experts and NASSCOM, found gaps in pedagogy, infrastructure and practical exposure to Generative AI, MLOps and foundational model development. The taskforce recommended raising practical exposure from the current 25–30% to between 40% and 75%, depending on the programme, with industry use cases from the first semester.
The timing makes this urgent. NASSCOM estimates India’s AI talent pool could grow from roughly 600,000–650,000 professionals to more than 1.25 million by 2027. The opportunity is expanding, but the bar to get into it is rising faster.
Adding an AI elective to a conventional computer science structure will not close that gap. Graduates need to have worked with real technologies, open-ended problems and ambiguity before their first job, because the job will no longer supply them. That belief shapes how we run Polaris School of Technology. PST has Microsoft, Google and Amazon Web Services as industry partners for their respective programmes. Students log more than 5,000 hours of hands-on coding, and internships begin in the first year itself.
Open source is where the idea shows most clearly. From their first weeks, students work on real software projects and contribute to code that developers around the world use. By third year they are building production systems. In just over two years, PST students have earned more than 36 open-source selections across highly selective and prestigious programs, including Linux Foundation, Google Summer of Code, C4GT, and more. They also learn from practising engineers, CTOs, founders and IIT alumni, with one-on-one mentorship built in. Through our academic ties with Grand Valley State Universities, San Diego State University and California State University, Fullerton, second-year PST students work alongside fourth-year students at these universities. That gives them an early measure of how their skills hold up alongside more experienced peers in another, global academic setting. These collaborations also open opportunities to spend a semester abroad and earn letters of recommendation from international faculty who have seen their work firsthand.
The principle behind these choices is simple: Responsibility develops through practice. A degree gives us four years to offer that practice, increase its difficulty and help students understand their mistakes.
PwC found that skills in the most AI-exposed jobs are changing more than twice as fast as in less-exposed roles, and that new tasks in them are 2.5 times more likely to rely on judgement, creativity and empathy. Technical skill still matters, but it now has to come with the ability to weigh AI output, handle uncertainty and make sound calls.
That makes the purpose of an engineering degree more demanding. India’s AI ambitions will depend less on how many engineers it produces than on whether they can contribute from the first day of a job that is being rewritten. The tools will keep changing and the expectations will keep rising. What lasts is the ability to learn, reason and build, and that now has to be taught before graduation.
The industry’s future senior engineers are sitting in classrooms today. The responsibility we give them there will shape the work they can take on tomorrow.
(The views expressed are personal)
This article is authored by Mukul Rustagi, founder, Classplus and Polaris School of Technology.