Skip to content

Repository files navigation

Eligify - Decision Automation Platform for Transparent Shortlisting

A rules-first, explainable decision automation platform that securely reads only relevant candidate data from existing systems and produces transparent shortlists.


📋 Project Overview

Eligify is a B2B SaaS platform designed to automate and transparentize the shortlisting and ranking of students or candidates for placements and hiring. Rather than replacing existing ERPs or HRMS systems, Eligify acts as a decision intelligence layer that:

  • Securely connects to existing data sources (Excel, ERP databases)
  • Fetches only relevant candidate data (final-year students, placement-opted candidates)
  • Applies rule-based eligibility and scoring logic
  • Produces ranked shortlists with clear explanations
  • Explains why each candidate was selected or rejected

🎯 Problem Being Solved

In India (and globally), colleges and companies face critical challenges during placements and hiring:

Problem Impact
Scattered Data Candidate data fragmented across Excel sheets and large ERPs
Manual Shortlisting Slow, error-prone, and inconsistent decisions
No Transparency HR teams cannot justify selection criteria
Diverse Criteria Each company has different eligibility requirements
No Audit Trail Compliance and decision history missing
Scalability Issues Manual processes break down with large candidate volumes

Core Insight: Decision-making does not scale with candidate volume.


🔑 Core Philosophy (Non-Negotiable)

1. Rules-First, Not AI-First

  • Eligibility rules always decide outcomes
  • AI will only assist later (not in MVP)
  • No black-box logic

2. Read-Only Integration

  • Never modify the organization's ERP
  • Never require schema changes
  • Minimal data access (fetch only required fields)

3. Explainability Over Automation

  • Every decision must be transparent
  • Rule-by-rule breakdown for each candidate
  • Audit trail for compliance

4. Low Adoption Friction

  • No SQL knowledge required from users
  • No ERP restructuring needed
  • Guided, familiar UX

✨ Key Features (MVP)

Data Ingestion

  • Excel / CSV Upload – Quick file-based data entry
  • Read-Only Database Integration – Secure connection to ERPs without schema changes
  • Column Mapping UI – Non-technical field mapping
  • Dataset Snapshots – Immutable, auditable data snapshots

Candidate Scope Filtering

  • Academic year / semester filters
  • Batch / passing year selection
  • Placement opt-in flags
  • Department / branch filters (optional)
  • Filters applied at source – Never fetch all students

Rule Engine

  • Mandatory Rules – Pass/fail eligibility criteria (e.g., CGPA ≥ 7)
  • Scoring Rules – Weighted scoring (skill match, internships, certifications)
  • Editable & Versioned – Change rules without data loss
  • Stored as Data – Not hardcoded

Evaluation & Ranking

  • Apply mandatory rules (exclude failing candidates)
  • Score eligible candidates
  • Sort by total score
  • Generate ranked shortlist
  • No randomness. No black box.

Explainability Dashboard

  • Rule-by-rule breakdown for each candidate
  • Score breakdown showing how points were calculated
  • Rejection reasons when candidates fail mandatory rules
  • Decision audit trail for compliance

Export & Integration

  • Export shortlisted candidates as CSV / Excel
  • Clean, guided user experience
  • Organization-based access control

🏗️ Architecture

Frontend (Next.js)
      ↓
Core Platform Backend (Node.js / NestJS)
      ↓
Dataset Snapshot + Rule Engine
      ↓
Ranked Shortlist + Explanations
      ↓
Export / View / Audit Trail

AI services will be added later as a separate layer.


💾 Tech Stack

Frontend

  • Framework: Next.js with TypeScript
  • Styling: Tailwind CSS
  • Components: shadcn/ui
  • Why: Professional UI, fast iteration, guided workflows

Backend (Core Platform)

  • Runtime: Node.js
  • Framework: NestJS with TypeScript
  • ORM: Prisma
  • Database: PostgreSQL
  • Why: Integration-heavy SaaS, clean architecture, strong typing, scalable APIs

Storage

  • File Uploads: AWS S3 or Azure Blob Storage
  • Structured Data: PostgreSQL

Authentication & Multi-Tenancy

  • Library: Auth.js or Clerk
  • Model: Organization-based multi-tenancy
  • Roles: Admin, Viewer

Deployment (MVP)

  • Frontend: Vercel
  • Backend: Render or Railway
  • Database: Managed PostgreSQL

📊 Data Ingestion Strategy

Option 1: Excel / CSV Upload (Fast Entry)

Perfect for demos, pilots, and small institutions.

User uploads spreadsheet
    ↓
Platform previews rows
    ↓
User maps columns (CGPA, skills, etc.)
    ↓
Data normalized and stored as snapshot

Option 2: Read-Only Database Integration (Key Differentiator)

For organizations already using ERPs.

Organization provides read-only DB credentials
    ↓
Platform discovers schema and lists tables
    ↓
User selects relevant table(s) and fields
    ↓
Platform applies filters at source:
  • Academic year = final year
  • Placement opted = true
  • Department = [selected]
    ↓
Filtered data fetched into snapshot
    ↓
DB connection closed
    ↓
No new tables created. No SQL written by org. No ERP modification.

🎯 Rule Engine (Core IP)

Rules are stored as data, making them editable, versioned, and auditable.

Rule Format

{
  "field": "cgpa",
  "operator": ">=",
  "value": 7.0,
  "type": "mandatory"
}

Rule Types

Type Example Outcome
Mandatory CGPA ≥ 7 Candidate passes or is excluded
Scoring Internship bonus (+10 points) Affects ranking, not eligibility

Evaluation Flow

  1. Apply mandatory rules – Fail or pass candidates
  2. Calculate scores – Weighted scoring for eligible candidates
  3. Sort by total score – Ranked shortlist generated
  4. Store evaluation result – Immutable record for audit

🔍 Explainability Example

Candidate: Aditya Sharma

Status: ✅ SELECTED (Rank #3)

Rule Breakdown:
  ✅ CGPA ≥ 7.0       [PASS] Current: 7.8
  ✅ No Active Backlogs [PASS] Backlogs: 0
  ✅ Has Required Skills [PASS] Python, SQL, Java
  ✅ Placement Opted [PASS]

Score Breakdown:
  Base Score: 50
  + Skill Match (Python, SQL): +15
  + Internship (6 months): +10
  + Certification (AWS): +5
  ───────────────────────
  Total Score: 80/100

Ranked #3 among 45 shortlisted candidates.

👥 User Experience Flow

  1. Create Organization – Multi-tenancy setup
  2. Upload File or Connect Database – Data ingestion
  3. Select Candidate Scope – Final year, placement opted, etc.
  4. Map Fields – Column/field mapping UI
  5. Create Rules – Define eligibility and scoring rules
  6. Run Shortlisting – Execute evaluation
  7. View Ranked Shortlist – Dashboard with rankings
  8. Click Candidate → See Explanation – Full rule breakdown
  9. Export Results – CSV / Excel download

UX Must Feel: Guided • Familiar • Safe • Transparent


📦 MVP Scope

✅ Included in MVP

  • Organization-based authentication
  • Candidate data ingestion (Excel / CSV & read-only DB)
  • Candidate scope filtering (academic year, placement-opted, batch)
  • Column and field mapping UI
  • Dataset snapshot creation
  • Rule-based eligibility engine
  • Score-based ranking
  • Explainable decision view
  • Export shortlisted candidates (CSV / Excel)
  • Clean, guided user experience

❌ Excluded from MVP

  • AI-driven decision making
  • Resume parsing
  • ERP write-back
  • Interview scheduling
  • ATS / HR workflows
  • Payroll / compliance features

🚀 AI / ML Layer (Planned, Not MVP)

AI will be added after the core platform is stable as a separate microservice.

Future Role of AI

  • Normalize skills and skill matching
  • Parse resumes for skill extraction
  • Suggest rule weightages and thresholds
  • Detect overly strict or loose criteria
  • Generate natural-language explanations
  • AI will never override rules

AI Tech Stack (Future)

  • Language: Python
  • Framework: FastAPI
  • Models: NLP, embeddings, skill classification
  • Architecture: Separate microservice

📈 Success Criteria for MVP

The MVP is successful if:

✅ Organizations can shortlist candidates in minutes, not hours
No ERP restructuring is required
Decisions are explainable to all stakeholders
Users trust the output and use it in production
Switching back to manual feels painful


🔮 Long-Term Vision

Eligify evolves into a decision intelligence layer for hiring:

  • Compliance-friendly screening system – Audit trails for regulatory bodies
  • Analytics engine – Identify skill gaps and trends
  • Trusted infrastructure – Fair, transparent, bias-reducing shortlisting

Future Use Cases

  • Campus placements
  • Lateral hiring
  • Scholarships
  • Admissions
  • Internal mobility
  • Contractor screening

⚠️ What NOT to Do

  • Do not build ATS features – Focus on shortlisting, not full HR workflows
  • Do not let AI decide eligibility – Rules decide outcomes, AI assists only
  • Do not require ERP schema changes – Read-only integration only
  • Do not overengineer infrastructure – MVPs must be lean
  • Do not chase features before trust – Explainability first

🛠️ Getting Started (Development)

Prerequisites

  • Node.js 18+
  • PostgreSQL 14+
  • npm or yarn

Installation (Coming Soon)

# Clone the repository
git clone <repo-url>

# Install dependencies
npm install

# Set up environment variables
cp .env.example .env.local

# Run database migrations
npm run db:migrate

# Start development server
npm run dev

📚 Documentation

  • PLAN.md – Detailed project specification
  • Architecture Guide – (To be added)
  • API Documentation – (To be added)
  • Database Schema – (To be added)

🤝 Contributing

(Contribution guidelines to be added)


📄 License

(License to be decided)


📞 Contact & Support

(Contact information to be added)


🎓 Project Context

Organization: College/Corporate Hiring
Region: India (Expanding Globally)
Problem Space: Campus Placements, Corporate Hiring, Admissions
Differentiator: Rules-First, Explainable, Read-Only Integration


Built with ❤️ to make hiring transparent, fair, and scalable.

About

Eligify is a B2B SaaS platform designed to automate and transparentize the shortlisting and ranking of students or candidates for placements and hiring. Rather than replacing existing ERPs or HRMS systems.

Resources

Contributing

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages