Skip to content
Value Creatives
All work
IKG Punjab Technical University logoGovernment

A Punjab government counselling engine that ranks and allots seats across 200+ colleges and multiple rounds — automatically, transparently, and provably at scale.

Government / Higher Education — Punjab, India

Book a callVisit IKG Punjab Technical University
The challenge

I.K. Gujral Punjab Technical University, a Punjab State Government body, runs centralized admission counselling across 200+ affiliated colleges and a dozen program families (B.Tech, B.Pharm, B.Arch, BHMCT, BBA, BCA, MBA, MCA, M.Tech, M.Pharm, Pharm.D and lateral entry). Ranking thousands of candidates and allotting state-quota seats across that network is a high-stakes, rules-dense process — separate merit formulas per program, reservation categories and sub-categories, Punjab vs other-state quota splits, TFW and minority buckets, age and percentage eligibility gates, and multi-round choice-filling with upgrades and refills. Done on the legacy stack it was slow, hard to audit, and error-prone, with little defensible trail for why any one candidate got any one seat — unacceptable for a government merit process where fairness must be provable.

Institutional architecture representing public-sector scale

Government

Government / Higher Education — Punjab, India

What we built

The system, in parts.

1

An automated rank-generation engine that reproduces each program's official merit formula in code: 10+2 aggregate for engineering, Physics+Chemistry+(Maths/Biology) for pharmacy, NATA+qualifying average for architecture, graduation percentage for postgraduate, and year-weighted diploma percentage for lateral entry — with per-program eligibility gates (minimum qualifying aggregate, minimum age on 31 December) and deterministic, reproducible tie-breaking

2

A deterministic seat-allotment engine that processes candidates strictly in merit order across the full network of colleges, honouring choice preferences, Punjab/other-state quota, reserved-category fallbacks, TFW and Sikh-minority seat buckets, and special-category and gallantry-award tie-break priority — so the same inputs always produce the same, explainable allotment

3

Multi-round counselling support: choice locking, round-over-round seat refills, seat upgrades that release a candidate's earlier seat back to the matrix, archival of consumed lower preferences, and carry-forward of unresolved choices into the next round, with admitted candidates excluded

4

Centralized seat-matrix management spanning 200+ institutes — frozen, snapshotted seat matrices per round so every allotment run is computed against an immutable, auditable seat picture

5

Three separate role-based portals (student, college, admin/superadmin) plus a shared Node API and typed domain layer, covering registration, document upload and verification, online fee payment (ICICI EazyPay) with SMS/email OTP, choice filling, provisional allotment letters, college PI reporting, direct/management-quota entry, eligibility verification, and Excel/PDF exports that preserve the university's existing column contracts

6

A full audit posture: every high-stakes action — rank generation, choice lock, seat-matrix freeze, allotment run, allotment publish, admission confirmation, withdrawal, correction — emits an audit event, and immutable snapshots of rank inputs, frozen matrices, and allotment input/output keep every decision explainable after the fact

Outcomes

What changed for them.

  • Manual, slow, error-prone counselling was replaced by an automated, repeatable engine that ranks candidates and allots seats across 200+ colleges and multiple rounds in minutes — at government admission scale

  • Seat allotment is transparent and merit-based by construction: candidates are served strictly in rank order against frozen seat matrices, with documented tie-break rules, so the university can defend exactly why each seat went where it did

  • Correctness was proven before go-live, not assumed — the engine was run against the real 2025-26 admission data and compared candidate-by-candidate against the university's own legacy allotments, with mismatches reported with plain-English reasons and a curated validation sample built for manual re-allotment cross-checking

  • Built for government-grade reliability and auditability: deterministic results, immutable per-round snapshots, and an audit event on every high-stakes operation give the university a defensible, reproducible record of the entire counselling process

  • Designed to absorb peak-admissions load with separate student/college/admin app surfaces behind a CDN, and to be re-runnable safely (idempotent allotment publish) so a round can be recomputed without corrupting prior state

How it’s built

The stack.

Next.js 15 (App Router)React 19TypeScript (strict, monorepo)Node.js APIMongoDB (transactions, indexed read models, immutable snapshots)Zod (typed contracts / validation)Deterministic domain engines (rank generation + seat allotment)ICICI EazyPay (payments)Government eSMS/DLT SMS + Office365 SMTP (OTP/notifications)AWS EC2 + S3 + CloudFront + Route 53 (PM2, golive/DNS scripts)Playwright + Vitest (legacy-parity & end-to-end validation)

More work

See all
Two ways to start

Want a system like this?

Tell us the workflow you want to run itself. We will scope a focused first project — designed, built, and operated, with humans in control.

Book a call
See our work