India produces too many engineers. An Expert Explains why it needs better ones

India’s engineering education system has expanded enormously, creating a vast pool of engineering talent and supporting the country’s rise as a technology and service economy. Access, once the principal challenge, has largely been achieved. The question now is quality — and whether the system is prepared for an era in which generative artificial intelligence (GenAI) is changing what students need to learn, how they learn, and how their learning should be assessed.

Why did expansion produce uneven outcomes?

India’s engineering education landscape changed dramatically after 2000. The Y2K-driven global demand for software professionals and the rapid growth of the IT industry created an unprecedented demand for engineering talent. India responded by expanding the network of Institutes of National Importance (INIs) — the Indian Institutes of Technology (IITs), National Institutes of Technology (NITs), and Indian Institutes of Information Technology (IIITs) — alongside thousands of public and private engineering institutions and colleges.

The objective was to expand access and produce the engineers needed by a growing economy. The expansion largely succeeded. Engineering education moved beyond metropolitan centres and elite institutions, reaching smaller towns and creating a vast national capacity.

But expansion produced widely varying outcomes across institutions. Premier INIs and a small group of other public and private institutions developed strong academic environments, faculty, research, industry links, and placements. Their continuing appeal shows that engineering itself has not lost its value.

At the same time, student demand has become increasingly concentrated in computer science, AI, data science, and allied disciplines. Many institutions have expanded seats in these areas, while traditional engineering branches struggle to attract students. The recently concluded JoSAA/CSAB Admission Counseling 2026 rounds have again highlighted this uneven demand, with many thousands of seats remaining vacant, including in parts of the premier system. Many institutions struggle to fill seats and remain viable, while others continue to expand popular programmes. Colleges that cannot attract students have increasingly faced progressive closure.

The paradox is not expansion, but expanding capacity without matching quality. Some institutions turned growth into excellence; many turned it into numbers. The result is a widening gap between producing capable engineers and merely awarding engineering degrees.

Where do India’s quality gaps lie?

India responded to the expansion with a regulatory and accreditation system. The All India Council for Technical Education (AICTE) prescribed norms and standards, while the National Board of Accreditation (NBA) introduced discipline-based accreditation, and National Assessment and Accreditation Council (NAAC) added institutional assessment. Yet regulation could not ensure quality.

Over time, compliance often became an end in itself. The faculty spent substantial time preparing reports, manuals, and documentary evidence. Laboratories, projects, and industry interaction were documented without demonstrating what students had actually learned. The gap between compliance and learning widened, at the cost of genuine teaching.

The classroom reflected this weakness. Rote learning and examination-driven assessment remained widespread. Curriculum changes often meant adding courses rather than changing how students learn. Laboratories and industry engagement too often became procedural exercises instead of opportunities to apply knowledge.

Faculty incentives further tilted the system towards what is easy to measure. Publications, citations, and other research metrics often carry greater weight than teaching innovation, mentoring, and student engagement. This encourages quantity over quality in mostly non-credible research, while teaching receives less institutional attention. 

Selected IITs, NITs, IIITs, and other strong institutions demonstrate that quality is achievable at scale. India has expanded engineering education, but quality has not kept pace with capacity. It has to move from documented compliance to demonstrated learning.

Why has GenAI changed the teaching equation?

GenAI has widened the gap between teaching and learning. Information, explanations, coding help, and project support are now instantly available through large language models (LLMs). When information is at every student’s fingertips, what exactly should an engineering teacher teach?

Yet much of engineering education continues with conventional lectures, assignments, and examinations, with little fundamental change in pedagogy. LLMs can explain concepts, write and debug code, generate assignments and project materials, assist with analysis, and support parts of research. The traditional advantage of the classroom — access to information — is therefore disappearing.

Rote learning and conventional lectures are increasingly inadequate, but curricula and assessment have been slow to respond. Pre-GenAI examinations cannot reliably establish what a student understands when answers, code, and project outputs can be AI-assisted. The traditional “teach, assign, and examine” model is therefore losing its ability to measure genuine learning.

The widening gap is ultimately between technology that has changed dramatically and pedagogy that has barely changed at all.

What should engineering education teach?

With GenAI, engineering education must move from information delivery to capability development. Assessment should focus less on recall and more on thinking, design, experimentation, problem-solving, and applying fundamentals to unfamiliar problems.

Projects should require students to frame problems, make engineering choices, test alternatives, and defend their solutions — not simply submit AI-generated outputs.

The teacher’s role must change from information provider to mentor, problem-framer, and learning facilitator. Engineers will not compete with LLMs in retrieving information or writing routine code. Their value will lie in understanding problems, questioning assumptions, making sound engineering judgements, working within real-world constraints, validating results, and taking responsibility for solutions.

The fundamental shift is from “What does the student know?” to “What can the student do with what they know and what they can access — and why?” AI can generate answers; engineering education must teach students to ask the right questions, challenge those answers,  and turn knowledge into reliable solutions.

What must change?

India has largely solved access; it must now solve learning. This requires continuous faculty development and stronger incentives for teaching and mentoring rather than documenting. Industry must become a genuine partner through live projects, internships, and curriculum input.

Student learning must be AI-enabled — use AI where it adds value, while developing what AI cannot replace: reasoning, design, experimentation, originality, and the ability to validate outcomes.

Institutions should be judged by demonstrated learning and outcomes, not seats, degrees, documents, or accreditation grades. The goal is simple: less compliance, more learning; fewer credentials, stronger engineers.

The author is a former computer science professor at IIT Kharagpur, IIT Kanpur, BITS Pilani, and JNU, and a former scientist at DRDO and DST.