Skip to content
UdgamThe Source

The Platform

How Udgam actually works.

For the technically curious — investors, chief technology officers and institutional IT heads doing due diligence. This page describes the architecture, the AI, the data boundaries and the integration surface.

  • Four-sided flywheel
  • Four AI engines
  • All data in India

How Udgam works

A four-sided flywheel, where every side must win.

This is the structural claim. If any one of the four sides puts in more than it gets back, that side disengages — and the whole system stops turning. Every design decision has to answer for its own side of this table.
Students

Gives

Engagement with their own outcomes, and activity data

Gets

Access tiers, opportunities, mentors, and a verified profile

Faculty

Gives

Course frameworks, assessment data, and approvals

Gets

AI assistance, public recognition, and income from industry

Institutions

Gives

Subscription and curriculum data

Gets

Accreditation reports, placement intelligence, and non-tuition revenue

Corporates & MSMEs

Gives

Hackathons, lectures, mentors, internships, and fees

Gets

Pre-verified talent, brand visibility, and research capacity

Why it compounds

Year 1

The platform has data. Outcomes, opportunities, engagements.

Year 2

Readiness scores are calibrated against actual placement results. The prediction becomes credible.

Year 3

Outcome-to-placement correlation is established — which learning outcomes produce which career results.

Year 5

The most comprehensive education-to-employment dataset in India. A data position no competitor can replicate by building faster.

The AI engines

Four engines, one principle: AI drafts, humans decide.

Nothing generated by a model reaches an official record without a human approval gate. That is not caution for its own sake — attainment data that cannot be defended to an accreditation reviewer is worse than no data at all.
01

Curriculum Intelligence

Faculty-facing

Reads a syllabus and drafts course outcomes, the outcome-to-programme matrix, question banks and rubrics — each at the cognitive level the material actually implies. Checks its own work for verb-to-level mismatches, and suggests corrective actions when a gap opens.

Faculty approve every draft. The head of department holds the gate on anything that affects an official record.

02

Learning Coach

Student-facing

Turns a student's record into a next action: which gap matters most, which practice sits at the right level, which opportunities actually fit, and what their trajectory looks like if nothing changes.

Nudges only. Nothing here alters a student's record.

03

Talent Intelligence

Corporate and MSME-facing

Translates an ordinary job description into the cognitive profile it implies, ranks candidates against it, drafts hackathon briefs and judging rubrics, and reports the translation back in the skills language recruiters already write in.

The company reviews and adjusts the translation before it is saved.

04

Platform Operations

Internal

Tags incoming listings, assembles accreditation reports from live data, flags assessment sets that lean too heavily on recall, detects anomalous patterns, and pre-screens listings for the human review team.

A person makes every listing decision. The engine only prepares the queue.

Security & data

Who owns what, and where it lives.

The first question every institutional IT head asks, and the first one an investor's diligence covers. Here are the answers without hedging.

Data sovereignty

  • All user data stored in India, in an India-region cloud
  • No transfer outside India except with explicit user consent
  • Compliant with the Digital Personal Data Protection Act 2023
  • Granular consent collected at registration, per data-use type, withdrawable at any time
  • Withdrawal triggers a deletion workflow within 30 days for personal data

Ownership boundaries

WhoOwns
StudentsTheir personal data and their own attainment records
InstitutionsProgramme data, faculty data, aggregate student data
Corporates and MSMEsTheir organisation and engagement data
UdgamPlatform usage analytics only — anonymised, no personally identifying data

Student protections

  • Personal data is never shared with a company unless the student applies to that listing
  • Attainment data reaches recruiters in translated skills form, not as raw percentages, unless the student opts in
  • Marks and grades are never visible to companies
  • Recruiter visibility can be switched off entirely, and is then respected absolutely — including in aggregate searches
  • No student data is ever sold to a third party

Institutional protections

  • Institutional attainment data is used for benchmarking only in fully anonymised aggregate form
  • Raw institutional data is never shared with another institution
  • Each institution sees its own position, never another institution's identity
  • Full export in CSV and Excel at any time
  • Exports remain available for two years after cancellation

Anti-spam, structurally

  • Companies cannot message a student who has not applied to them
  • No cold outreach is possible — engagement is student-initiated or institution-approved
  • A hard limit of three automated notifications per user per day
  • Every listing is reviewed by a person before it goes live
  • Rejected listings can be appealed, and are re-reviewed within 48 hours

ERP & LMS integration

Zero data re-entry. This is the whole objective.

Student marks, enrolment and course data already exist in your institution's systems. Anything Udgam asks a faculty member to re-type is a design failure.
IntegrationDirectionMethod
Student data syncERP → Udgam — enrolment and roll numbersREST API or CSV upload
Marks importERP → Udgam — assessment scoresREST API or CSV template
Course data syncERP → Udgam — course list and examination schemeREST API or CSV
Attendance syncERP → Udgam — optional, as an indirect proxyREST API
Outcome report exportUdgam → ERP — attainment for ERP reportingREST API or CSV
Alumni recordsERP → Udgam — graduation recordsCSV bulk upload

Connector architecture

  • OAuth 2.0 authentication
  • Rate-limited API — up to 10,000 records per sync batch
  • Error logging with retry logic
  • Sync frequency configurable by your IT team: daily, weekly or on demand
  • Documented REST API for any custom or in-house ERP

Reaching students who are not on good connections

  • Progressive web app — installable on mobile without an app store
  • Offline: last-synced outcomes, downloaded assessments, drafted reflections that sync on reconnect
  • Low-bandwidth mode reduces data transfer by around 80%
  • Text-only mode for 2G connections
  • Critical alerts delivered by WhatsApp even on 2G
  • Six languages at launch: English, Hindi, Marathi, Tamil, Telugu and Kannada

The Bloom's framework

Explained for people who do not work in education.

This is the single concept that makes Udgam's talent signal different from every other platform. It is worth five minutes even if you never intend to use the word.

Bloom’s Taxonomy is a fifty-year-old framework that sorts thinking into six levels, from recalling a fact up to creating something that did not previously exist. It is the standard vocabulary of curriculum design worldwide, and Indian accreditation has required its use since 2013.

What nobody had done before is connect it to hiring. A degree certificate tells an employer that someone completed a course. It says nothing about whether that person can analyse an unfamiliar problem or only reproduce a familiar one. Those are different capabilities, they are separately measurable, and the difference is most of what separates a productive graduate from an expensive one.

LevelWhat it means in plain termsTypically hired for
1 — RememberRecall a fact you were taughtRarely a hiring signal on its own
2 — UnderstandExplain it in your own wordsSupport and coordination roles
3 — ApplyUse a known method on a familiar problemIT services and execution-heavy work
4 — AnalyseBreak an unfamiliar problem apart and find its structureProduct engineering, systems, data
5 — EvaluateJudge between competing approaches and defend the judgementSenior individual contributors, consulting
6 — CreateProduce something that did not exist beforeResearch, innovation, founding teams

Why the weighting matters

Higher-order attainment is weighted substantially more heavily than recall. A student who performs well at Create level scores above one who performs near-perfectly at Remember level. That is not an aesthetic preference — it reflects hiring reality, and it makes the tier system structurally resistant to gaming. Memorisation cannot unlock a premium opportunity.

Why recruiters never see the word

A translation layer converts attainment data into the skills language HR already writes in. A recruiter reads “strong analytical problem-solving”; the matching engine underneath is operating on cognitive level. The framework is the machinery, not the interface.

Explore the levels

Click through the six.

Each one with a plain-English definition, what it looks like in a student, and which roles require it.

Bloom's six levels

Higher levels are weighted disproportionately — Create-level attainment counts six times a Remember-level pass. That is what makes the tier system resistant to rote learning.

Level 3

Apply

Use a known method on a familiar problem.

What it looks like in a student

Given a problem of a type you have seen before, you can pick the right approach and execute it reliably.

Which roles need it

IT services and execution-heavy roles. The largest single category of graduate hiring in India.

The stack

What it is built on.

Web and mobile

A single codebase across web, iOS and Android, installable as a progressive web app with genuine offline support for low-connectivity campuses.

Data core

A relational core for outcome hierarchies and attainment records, where auditability and referential integrity matter, with separate stores tuned for feeds and sessions.

AI layer

Frontier language models for drafting and generation, with semantic search and a hybrid recommendation engine for matching. Model choice is kept swappable by design.

Search

A dedicated search cluster for talent, opportunities, the faculty directory and the innovation project catalogue.

Infrastructure

India-region cloud, mandatory for DPDP compliance. Containerised services, continuous deployment, error tracking, and a 99.5% uptime commitment to institutions.

Payments

Indian payment rails — UPI, cards, net banking, and purchase orders for institutions paying from government grants. Escrow for large prize pools and consulting milestones.

99.5%

uptime, committed contractually to subscribing institutions

6

languages at launch, covering the platform's target state markets

India

where every byte of user data is stored, without exception

AI generates outcome frameworks, question banks, rubrics, attainment summaries and matching recommendations. Humans review and approve before anything affects an official record. AI is a first-draft engine; authority remains human.
Udgam PRD — Design Principle 4
  • 10X Faster

    Accreditation in a click. Projects in a semester. Placements in a season.

  • Talent Access

    India's verified talent — from Nagpur, Coimbatore, Guwahati, not just IITs.

  • Co-working

    Universities and industry, working on the same problems at the same time.

Due diligence

Ask us the architecture questions.

If you are an institutional IT head, a CTO or an investor, we would rather answer the hard version properly than the easy version quickly.