The default that became a trap

Walk into any JEE coaching centre in India in 2026. Ask the Class 11–12 students what branch they want. ~70% will say "CSE" — Computer Science and Engineering.

Ten years ago this signalled aspirational ambition. Today it signals lack of choice. CSE has become the default that most engineering aspirants reach for without examining whether it actually fits them. The cost shows up at graduation: ~48% of Indian engineering graduates are unemployable in their chosen stream, per Mercer-Mettl's annual India Skills Report. CSE specifically — once the safest engineering bet — now produces a significant fraction of those unemployable graduates because the supply has outpaced quality demand.

This piece is the critical look. Not "CSE is bad" — it isn't. It's "CSE is the wrong choice for most of the 70% reaching for it." Here's why.

Reality check 1: saturation by the numbers

India produces approximately 1.5 million engineering graduates per year. CSE accounts for roughly 40–45% of that intake. That's 600,000+ CSE graduates entering the job market annually.

The actual hiring at the entry-level for genuine software engineering roles in India runs in the range of 200,000–300,000 per year, including consulting and IT services. The mathematical gap is real.

The gap closes via three mechanisms:

  • Tier-1 graduates (IIT, IIIT, BITS, top NITs) get hired at full software-engineer salaries. Maybe 50,000 graduates per year.
  • Tier-2 graduates (state engineering colleges, mid-tier private universities) get hired into IT services (TCS, Infosys, Wipro, Cognizant) at substantially lower compensation. ~150,000–200,000 per year.
  • Tier-3 and below — this is where the unemployable cohort sits. Fresh CSE graduates working in non-engineering roles, returning home, or sitting unemployed.

If you're not on track for tier-1 college admission, the CSE economic outcome you're imagining (₹15–₹40 LPA out of college) is not the outcome you'll actually get.

Reality check 2: salary stratification

The "₹40 LPA package" headlines from IIT placement seasons obscure the median outcome.

| College tier | Realistic CSE entry CTC range | | :-- | :-- | | IIT (Top 5) | ₹15–₹40 LPA (top quartile much higher) | | IIT (other) + IIIT-Hyderabad + BITS | ₹12–₹25 LPA | | Top NITs | ₹8–₹18 LPA | | State Tier-1 colleges (Anna, COEP, RV, PESIT, VIT, etc.) | ₹6–₹12 LPA | | State Tier-2 colleges | ₹3.5–₹6 LPA (largely IT services) | | Tier-3 colleges | ₹1.8–₹4 LPA (if placed at all) |

The compounding effect over a career narrows the gap somewhat, but the starting tier sets a 5–7 year ceiling that most graduates don't escape without pivoting through MBA, Master's abroad, or independent skill-building.

If your Class 11–12 academic trajectory points to a tier-3 engineering college, the realistic CSE outcome is a ₹3 LPA IT services job. That's a structurally different decision than the ₹20 LPA IIT-CSE outcome — and it should be evaluated on different terms.

Reality check 3: AI is changing entry-level demand

The 2024–2026 hiring slowdown at major Indian tech firms is partly cyclical, partly structural. Generative AI has compressed the labour requirement for entry-level coding tasks. TCS, Infosys, Wipro, Cognizant, HCL — the mass employers of CSE graduates — have all reported structurally lower entry-level intake for FY 2024–25 and FY 2025–26.

The roles AI is displacing fastest are the entry-level "do exactly what the spec says" coding jobs that historically absorbed Indian IT services graduates. Senior engineers, system designers, product engineers, and AI specialists are still in demand — but the entry path is increasingly squeezed.

For students entering CSE in 2026, the assumption "I'll get a ₹6 LPA IT services job" is a much weaker assumption than it was for 2018 graduates. Plan accordingly.

Reality check 4: most CSE programmes don't teach real engineering

Walk through the curricula of mid-tier Indian engineering colleges. Compare to what's actually demanded in industry. The gap is large:

  • Most colleges teach C, C++, Java at intro level — useful but not what most modern jobs use.
  • Few teach modern web frameworks (React, Node, Django, Rails), cloud platforms (AWS, GCP, Azure), DevOps, or production engineering practices.
  • Most don't have working software engineering practice — version control, testing, code review, CI/CD — taught as part of the core curriculum.
  • Few have real-world capstone projects with industry partners. The standard "final-year project" is a small academic exercise, not industry-grade work.

The graduates who land good jobs are those who self-taught what their college didn't teach — through online courses, open-source contributions, hackathons, and internships. The college credential alone is increasingly insufficient.

When CSE genuinely fits

Despite all of this, CSE remains a strong choice for the right student. The right student profile:

  1. Genuine intellectual interest in computing. Not "I want a stable job" but "I find programming and systems thinking actually engaging." If you've been programming for fun (not just for school assignments) before Class 12, that's a real signal.
  2. Capacity for self-directed learning. Modern software engineering changes faster than any college can teach. The graduates who thrive are the ones who learn from documentation, blog posts, open source, and side projects — not just classes.
  3. Tier-1 or strong-tier-2 college admission realistic. If your trajectory points to top NITs, IIITs, top state colleges, or above, CSE remains a high-leverage choice. Below that tier, alternatives may serve you better.
  4. Comfort with rapid technological change. What's hot in 2026 won't be hot in 2030. Comfort with ambiguity and continuous skill rebuilding is a personality trait, not a learned skill.

If three of these four are true for you, CSE is genuine. If two or fewer, you're likely defaulting.

Alternatives to consider

If CSE is the default but the fit is uncertain, these alternatives often produce better outcomes for the right student:

  • Electronics and Communication Engineering (ECE) — broader career options, less saturated, transitions into software roles work fine for those who later want it.
  • Mechanical Engineering — undervalued in current market sentiment but core to manufacturing, EV, aerospace, and robotics. Lower starting compensation but stronger long-term resilience for technically-inclined students.
  • Civil Engineering — large infrastructure pipeline (highways, metros, smart cities). Stable but specialised.
  • B.Tech in Mathematics and Computing or AI/ML specialisations at top schools — captures the genuine quantitative depth without the CSE crowd.
  • B.Sc. Statistics or Mathematics + M.Sc. — for genuinely quantitative students, may produce stronger AI/ML and quant-finance pathways than B.Tech CSE at non-elite colleges.
  • Integrated programmes at IIITs and IISERs — narrower entry, stronger research-aligned path.
  • Direct B.Sc. → M.Sc. CS or Data Science from top universities — bypasses the engineering-college framework entirely, often produces equivalent industry outcomes.

The right alternative depends on your specific aptitude profile and academic strength.

How to decide honestly

Three questions:

  1. What's your honest college admission probability? Be specific — not "I'll try for IIT" but "given my Class 11 mock test scores, what realistic colleges am I looking at?" This determines which CSE outcome you're actually buying.
  2. Is your interest in computing genuine or social? Have you written code that wasn't assigned? Built anything you didn't have to? If yes, CSE genuinely fits. If no, the interest is likely social signal, not real attraction.
  3. What's your behavioural profile? CSE rewards high-Diligence + high-Patience profiles (sustained focus on hard problems, careful debugging). If your RAPD profile leans high-Affiliation or high-Results, fields like product management, sales engineering, or solutions consulting may use your strengths better than software engineering itself.

How Dheya helps with engineering branch selection

Most engineering branch decisions in Indian families happen on a 10-day timeline after JEE results. The pressure makes default decisions inevitable.

Dheya's structured assessment provides three things that the 10-day window doesn't:

  1. A behavioural profile match against the day-to-day work of each branch — not the salary, not the headline outcomes, but what an engineer actually does in CSE vs Mechanical vs ECE vs Civil.
  2. A college-tier realistic outcome map — given your scores, what's the actual outcome distribution at colleges you can realistically attend?
  3. A parent-included conversation — most engineering decisions involve family pressure. Having structured data in front of everyone re-frames the conversation from "Science vs not-Science" to "what's actually best for this specific student."

The output is rarely "CSE yes/no." More often it's "CSE at a tier-1 college fits you; CSE at a tier-3 college does not — here's why, and here's an alternative that does fit at the colleges you can attend."

FAQs

Is CSE still worth it in 2026?

Yes, for tier-1 college admission with genuine interest in computing. No, for tier-3 admission with no genuine interest. The middle case (tier-2 college, moderate interest) requires harder honest analysis than most families do.

What if I've already committed to CSE at a non-tier-1 college?

Optimise the next 4 years. (a) Build a portfolio (real projects, GitHub contributions, side products). (b) Compete in hackathons and coding contests (Codeforces rating ≥1900 opens doors that college brand doesn't). (c) Intern at small startups and product companies, not just service firms. (d) Specialise in something defensible — AI/ML, security, distributed systems, embedded — rather than generic full-stack. (e) Plan for a master's abroad or a domestic strong M.Tech if entry placements are weak.

Is CSE better than ECE?

Different optimisations. CSE is the saturated default; ECE is less crowded with stronger career resilience for hardware-software hybrid roles. Both can lead to software roles. ECE keeps more options open if computing turns out not to fit; CSE doesn't.

What about Data Science / AI / Machine Learning specialisations?

If the college's "AI/ML specialisation" is just rebranded CSE, it's the same as CSE. If the programme has genuine Maths and Statistics depth (linear algebra, probability, statistical learning theory), strong faculty in active research, and access to GPU compute, it's a different programme. Verify before assuming.

Should I do CSE just because everyone is doing it?

This is the worst reason. The 70% who chose CSE because everyone else did are the same 70% staring at modest IT services job offers in 2026. The right reason to do CSE is genuine interest plus realistic college tier.

What if my parents insist on CSE?

This is a real conversation, not a yes/no. Most parents care about the outcome (stable career, good income), not the label CSE. Showing structured data — your realistic college outcomes, alternatives that match your aptitude better, the actual employment statistics for graduates from colleges in your reach — usually shifts the discussion. Dheya's parent-included sessions are designed exactly for this conversation.