SkillBridge — AI Adaptive Onboarding Engine

Stop wasting time teaching what people already know.

SkillBridge is an agentic AI-powered onboarding engine that analyzes a new hire's resume against a job description, identifies exact skill gaps, and generates a personalized learning roadmap — delivered instantly to their inbox.

🔗 Live Demo: https://skillbridge-ui.onrender.com
📹 Demo Video: https://drive.google.com/file/d/1RFxNGo4UqrKA59YxxLgq-st8p5EptGBC/view?usp=sharing

📂 GitHub: https://github.com/shahanakarthikeyan/skillbridge

---

Features

• 📄 PDF Resume Upload — extracts and parses resume text automatically
• 🤖 Agentic AI Pipeline — Groq (Llama 3.1) with tool-calling for structured analysis
• 📊 Skill Gap Analysis — detects missing skills and proficiency level gaps
• 🎯 Proficiency-Aware Matching — identifies not just WHAT is missing but HOW BIG the gap is
• 🗺️ Personalized Learning Roadmap — prioritized courses mapped to real skill gaps
• 🧠 Reasoning Trace — full step-by-step agent explanation of every recommendation
• 📧 Email Export — sends the roadmap as a PDF to the user's inbox via Gmail
• 🌐 Cross-Domain — works across Tech, Cloud, Data, UI/UX, Operations and more

---

Tech Stack

Layer	Technology
Frontend	React + Vite + Tailwind CSS
Backend	FastAPI + Python
AI / LLM	Groq API — Llama 3.1
PDF Parsing	PyMuPDF
Email	Gmail SMTP
Deployment	Render

---

How It Works

Resume PDF + Job Description
          ↓
     React Frontend
          ↓
    FastAPI Backend
          ↓
  Groq Agent (Llama 3.1)
    ├── extract_skills()       → 28 skills from resume, 5 from JD
    ├── compute_skill_gap()    → coverage 80%, 1 critical gap found
    └── build_learning_pathway() → 1 module, ~10h roadmap
          ↓
 Learning Roadmap + Reasoning Trace
          ↓
  React UI Display + Gmail Email Export

---

Skill Gap Analysis Logic

Step 1 — Skill Extraction

The Groq agent calls extract_skills() on both resume and JD and returns:

• skill_name — e.g. Python, AWS, SQL
• current_level — none / beginner / intermediate / advanced
• required_level — from the job description
• gap_severity — critical / moderate / minor

Step 2 — Gap Algorithm

FOR each required skill in JD:

  IF skill missing in resume        → 🔴 CRITICAL GAP (high priority)
  IF current_level < required_level → 🟡 PARTIAL GAP  (medium priority)
  IF current_level >= required_level → ✅ MATCHED

Coverage % = (Matched Skills ÷ Total JD Skills) × 100

Step 3 — Adaptive Path Selection

• Gaps ranked by severity (critical first)
• Each gap mapped to best-fit course from a fixed catalog
• Course selected based on: skill name + current level + target level
• Output: ordered roadmap, most urgent course first

Step 4 — Reasoning Trace

The agent generates a natural language conclusion:

"The individual has a strong technical background with expertise in Python and Machine Learning. However, there is a critical gap in AWS — required by the JD at beginner level but currently at none. Recommended pathway: AWS Cloud Practitioner certification."

---

Setup & Installation

Prerequisites

• Python 3.9+
• Node.js 18+
• Groq API Key — free at console.groq.com
• Gmail App Password

Backend

cd backend
pip install -r requirements.txt

Create a .env file:

GROQ_API_KEY=your-groq-api-key
GMAIL_USER=your@gmail.com
GMAIL_APP_PASSWORD=your-app-password

Run the backend:

uvicorn main:app --reload

Frontend

cd frontend
npm install
npm run dev

App runs at: http://localhost:3000

---

Datasets Referenced

Dataset	Source	Role
Resume Dataset	Kaggle — Sneha Anbhawal	Skill taxonomy inspiration
O*NET Database	onetcenter.org	Job role competency reference
Job Descriptions	Kaggle — Kshitiz Regmi	JD pattern reference

Note: SkillBridge does not train on external datasets. Skill extraction is performed dynamically by Groq (Llama 3.1) on user-provided resume and JD text. The above datasets were referenced for skill taxonomy and domain understanding.

---

Metrics

Metric	Result
Skill extraction (resume)	28 skills detected
Skill extraction (JD)	5 skills detected
Coverage accuracy	80% on test case
Proficiency level detection	none / beginner / intermediate / advanced
Roadmap generation time	< 10 seconds
Hallucination rate	0% (catalog-grounded)
Email delivery	✅ Confirmed working

---

Project Structure

skillbridge/
├── frontend/
│   ├── src/
│   │   ├── components/
│   │   └── App.jsx
│   ├── package.json
│   └── vite.config.js
├── backend/
│   ├── main.py
│   ├── agent.py
│   ├── requirements.txt
│   └── .env
└── README.md

---

Team

K. Shahana — shahanakarthikeyan0@gmail.com
R. Rajeshwari

---

Built For

ARTPARK CodeForge Hackathon — AI Adaptive Onboarding Engine Challenge

---

License

MIT License