🎯 CareerPath AI — Career Guidance System

An AI-powered career guidance system that provides personalized career predictions and roadmaps for students who just passed 10th or 12th standard. Built with a Python Flask backend (ML-powered) and a modern React frontend.

📸 Screenshots

Landing Page

Premium dark-themed landing page with animated hero section and feature highlights.

Profile Assessment

Interactive profile page with score sliders, skill tags, and interest selection.

AI Career Predictions

Dashboard showing top 3 career recommendations with confidence scores from the ML ensemble.

Personalized Roadmap

Step-by-step career roadmap tailored to the student's education level (after 10th or 12th).

🏗️ System Architecture

┌─────────────────────────────┐
│     User Interface (React)  │
└─────────────┬───────────────┘
              ▼
┌─────────────────────────────┐
│   Input Processing Module   │
│  (Scores, Skills, Interest) │
└─────────────┬───────────────┘
              ▼
┌─────────────────────────────┐
│  Feature Engineering &      │
│  Preprocessing              │
└──────┬──────────────┬───────┘
       ▼              ▼
┌──────────────┐ ┌──────────────┐
│ Random Forest│ │   XGBoost    │
│  Classifier  │ │  Classifier  │
└──────┬───────┘ └──────┬───────┘
       └────────┬───────┘
                ▼
┌─────────────────────────────┐
│  Prediction & Recommendation│
│  (Weighted Ensemble)        │
└─────────────┬───────────────┘
              ▼
┌─────────────────────────────┐
│   Output — Career Results   │
│  Suggestions, Skills, etc.  │
└─────────────────────────────┘

🔄 Workflow

Step	Process	Description
1	User Registration	Student creates account with education level (10th/12th)
2	Skill & Interest Analysis	System collects academic scores, skills, and interests
3	Career Domain Prediction	ML model predicts suitable career domains
4	Career Mapping	System maps careers to required skills and courses
5	Roadmap Generation	Personalized career roadmap is created
6	Dashboard Display	Career options, certifications, and timeline displayed

🛠️ Tech Stack

Backend

Technology	Purpose
Python 3.10+	Backend language
Flask	REST API framework
scikit-learn	Random Forest classifier
XGBoost	XGBoost classifier
PyJWT	JWT authentication
Flask-CORS	Cross-origin support
NumPy	Numerical computing

Frontend

Technology	Purpose
React 19	UI library
Vite	Build tool & dev server
React Router	Client-side routing
Vanilla CSS	Custom design system
Google Fonts (Inter)	Typography

📁 Project Structure

CareerPathAi/
├── backend/
│   ├── app.py                  # Flask entry point
│   ├── requirements.txt        # Python dependencies
│   ├── models/
│   │   ├── career_data.py      # Career datasets, roadmaps, training data
│   │   └── ml_model.py         # RF + XGBoost ensemble predictor
│   └── routes/
│       ├── auth.py             # Register/Login (JWT)
│       ├── predict.py          # Career prediction endpoint
│       └── roadmap.py          # Roadmap generation endpoint
│
├── frontend/
│   ├── index.html              # Entry HTML
│   ├── package.json            # Node dependencies
│   ├── vite.config.js          # Vite configuration
│   └── src/
│       ├── main.jsx            # React entry point
│       ├── App.jsx             # Router & auth state
│       ├── api.js              # Backend API client
│       ├── index.css           # Design system (dark theme)
│       ├── components/
│       │   └── Navbar.jsx      # Navigation bar
│       └── pages/
│           ├── Landing.jsx     # Hero & features page
│           ├── Auth.jsx        # Login/Register page
│           ├── Profile.jsx     # Score/Skill/Interest input
│           ├── Dashboard.jsx   # Career prediction results
│           └── Roadmap.jsx     # Personalized career roadmap
│
└── .gitignore

🚀 Getting Started

Prerequisites

Python 3.10+ installed
Node.js 18+ and npm installed
Git installed

1. Clone the Repository

git clone https://github.com/atharvpatil01/CareerPathAi.git
cd CareerPathAi

2. Setup Backend

cd backend
pip install -r requirements.txt
python app.py

The backend will:

Generate synthetic training data (2400 samples)
Train Random Forest + XGBoost models
Start the API server on http://localhost:5000

3. Setup Frontend

cd frontend
npm install
npm run dev

The frontend will start on http://localhost:5173

4. Open the App

Navigate to http://localhost:5173 in your browser and start exploring!

📡 API Endpoints

Method	Endpoint	Description
`POST`	`/api/register`	Register new user (name, email, password, education_level)
`POST`	`/api/login`	Login with email & password
`POST`	`/api/predict`	Get career predictions (scores, skills, interests)
`GET`	`/api/options`	Get available skills & interests
`GET`	`/api/roadmap/<career>`	Get career roadmap (query: education_level)
`GET`	`/api/careers`	List all available careers
`GET`	`/api/health`	Health check

Example: Predict Career

curl -X POST http://localhost:5000/api/predict \
  -H "Content-Type: application/json" \
  -d '{
    "scores": {"math": 85, "science": 78, "english": 70, "logical_reasoning": 90},
    "skills": ["Programming", "Problem Solving", "Analytical Thinking"],
    "interests": ["Technology", "Science", "Engineering"]
  }'

🤖 ML Model Details

Ensemble Architecture

The system uses a weighted ensemble of two classifiers:

Random Forest (150 estimators, max_depth=15) — Weight: 45%
XGBoost (150 estimators, max_depth=8, lr=0.1) — Weight: 55%

Training Data

2400 synthetic samples generated from career profiles
34 features: 4 academic scores + 15 skill indicators + 15 interest indicators
12 career domains with distinct score/skill/interest patterns

Career Domains

Domain	Key Strengths
💻 Software Engineering	High math + logical reasoning + programming
📊 Data Science	High math + analytical thinking + data analysis
🏥 Medicine	High science + research + healthcare interest
⚖️ Law	High english + communication + critical thinking
📈 Business & Finance	High math + leadership + business interest
🎨 Design & UX	Creativity + design skills + arts interest
👩‍🏫 Teaching & Education	Communication + public speaking + education interest
⚙️ Mechanical Engineering	High math + science + engineering interest
🏗️ Civil Engineering	High math + science + environment interest
📰 Journalism & Media	High english + writing + media interest
🧠 Psychology & Counseling	Communication + empathy + social work interest
🧬 Biotechnology	High science + research + healthcare interest

🗺️ Personalized Roadmaps

Each career provides a tailored roadmap based on whether the student passed 10th or 12th standard, including:

📚 Step-by-step educational path (stream selection, degree, specialization)
🛠️ Skills to develop (technical + soft skills)
📜 Certifications to pursue
🎓 Top courses & colleges recommended
💰 Expected salary range in India

✨ Features

🔐 JWT Authentication — Secure register/login with token-based auth
📊 Interactive Score Input — Slider-based academic score entry
🏷️ Tag Selection — Click-to-select skills and interests
🤖 AI Predictions — Top 3 career matches with confidence percentages
🗺️ Visual Roadmap — Animated timeline with career milestones
🌙 Premium Dark Theme — Glassmorphism, gradients, micro-animations
📱 Responsive Design — Works on desktop, tablet, and mobile
⚡ Fast & Modern — Vite-powered frontend with hot reload

🧪 Testing

Backend API Test

# Health check
curl http://localhost:5000/api/health

# Register
curl -X POST http://localhost:5000/api/register \
  -H "Content-Type: application/json" \
  -d '{"name":"Test","email":"[email protected]","password":"1234","education_level":"12th"}'

# Get options
curl http://localhost:5000/api/options

Full Flow Test

Open http://localhost:5173
Register with your details (select 10th or 12th)
Fill in academic scores, select skills & interests
Click "Get Career Predictions"
View your top 3 career matches
Click "View Career Roadmap" for detailed guidance

📄 License

This project is open source and available under the MIT License.

👤 Author

Atharv Patil

GitHub: @atharvpatil01

Made with ❤️ for students navigating their career paths

📋 Project Methodology

1. Introduction & Problem Statement

Problem: Students who have just passed their 10th or 12th standard face a critical decision about their career path but often lack personalised, data-driven guidance.

Solution: An AI-powered web application that takes a student's academic scores, skills, and interests as input and uses Machine Learning to predict the top 3 suitable career domains — along with a personalised step-by-step roadmap.

2. Development Methodology

The project follows a modified Agile/Iterative approach split into clearly defined phases:

Phase 1 — Requirement Analysis & Data Modelling

Activity	Output
Identify target audience (10th/12th pass students)	User persona document
Map 12 career domains with score ranges, skills, and interests	`career_data.py` — `CAREER_PROFILES` dictionary
Define input features (4 scores + 15 skills + 15 interests = 34 features)	Feature schema
Design career roadmaps for each domain (after 10th & after 12th)	`CAREER_ROADMAPS` dictionary

Phase 2 — Machine Learning Model Development

Activity	Output
Generate synthetic training data (2,400 samples) using CAREER_PROFILES	`generate_training_data()` in `career_data.py`
Train Random Forest (150 estimators, max_depth=15)	Trained RF model in `ml_model.py`
Train XGBoost (150 estimators, max_depth=8, lr=0.1)	Trained XGB model in `ml_model.py`
Build weighted ensemble (RF 45% + XGBoost 55%)	`CareerPredictor.predict()` method
Validate training accuracy	Accuracy printed on startup

Phase 3 — Backend API Development (Flask)

Activity	Output
Create Flask app with CORS support	`app.py` — `create_app()`
Implement JWT-based auth (register/login)	`routes/auth.py`
Build prediction endpoint	`routes/predict.py` — `POST /api/predict`
Build roadmap generation endpoint	`routes/roadmap.py` — `GET /api/roadmap/<career>`
Add utility endpoints (health, options, careers)	Various routes

Phase 4 — Frontend Development (React + Vite)

Activity	Output
Design dark-themed design system (CSS variables, glassmorphism)	`index.css` — 22 KB design system
Build responsive Navbar with auth awareness	`Navbar.jsx`
Build Landing page with hero section & features	`Landing.jsx`
Build Auth page (login/register toggle)	`Auth.jsx`
Build Profile page (sliders for scores, tag-select for skills/interests)	`Profile.jsx`
Build Dashboard (top 3 predictions with confidence bars)	`Dashboard.jsx`
Build Roadmap page (animated timeline)	`Roadmap.jsx`
Wire up API client	`api.js`

Phase 5 — Integration & Testing

Activity	Output
Connect frontend → backend via REST API	`api.js` baseURL → `localhost:5000`
End-to-end flow testing (register → login → profile → predict → roadmap)	Manual test script in README
API endpoint testing with `curl`	Test commands documented

3. Technology Stack Justification

Layer	Choice	Rationale
ML — Random Forest	scikit-learn	Robust, interpretable, handles tabular data well
ML — XGBoost	xgboost	Best-in-class gradient boosting for structured data
Ensemble Strategy	Weighted average (45/55)	Combines RF's stability with XGBoost's precision
Backend	Flask	Lightweight, Pythonic, easy ML integration
Auth	PyJWT	Stateless token-based auth, no DB dependency
Frontend	React 19 + Vite	Fast dev server, modern JSX, component reusability
Styling	Vanilla CSS	Full control, no framework overhead, premium dark theme
Routing	React Router	SPA navigation, auth-protected routes

4. Data Flow — Complete User Journey

Student                    React Frontend              Flask Backend               ML Ensemble
  │                              │                           │                          │
  ├── Register ────────────────► POST /api/register ────────►│                          │
  │◄──────────── JWT Token ◄──── JWT Token ◄────────────────┤                          │
  │                              │                           │                          │
  ├── Input Scores/Skills ─────► POST /api/predict ─────────► prepare_features()        │
  │                              │                           ├── 34-dim vector ────────►│
  │                              │                           │   RF predict_proba       │
  │                              │                           │   XGB predict_proba      │
  │                              │                           │◄── Ensemble (0.45*RF     │
  │                              │                           │     + 0.55*XGB)  ◄───────┤
  │◄──── Dashboard (3 cards) ◄── Top 3 careers + confidence◄┤                          │
  │                              │                           │                          │
  ├── Click "View Roadmap" ────► GET /api/roadmap/<career> ─►│                          │
  │◄──── Animated Timeline  ◄─── Roadmap phases, skills ◄───┤                          │

5. ML Model Methodology

5.1 Synthetic Data Generation

Since real student-career outcome data is not available, the system generates 2,400 synthetic training samples (200 per career domain) using the curated CAREER_PROFILES:

Score features (4): Gaussian distribution centred on each career's ideal range (e.g., Software Engineering → math 75–100, logical 80–100)
Skill features (15): Binary flags, 75% probability if the skill is associated with the career, 15% otherwise
Interest features (15): Binary flags, 75% probability if associated, 12% otherwise

5.2 Model Training

Model	Hyperparameters
Random Forest	`n_estimators=150`, `max_depth=15`, `min_samples_split=5`, `min_samples_leaf=2`
XGBoost	`n_estimators=150`, `max_depth=8`, `learning_rate=0.1`, `subsample=0.8`, `colsample_bytree=0.8`

5.3 Ensemble Prediction

ensemble_probabilities = 0.45 × RF_probabilities + 0.55 × XGB_probabilities

Top 3 indices are extracted via argsort(), inverse-label-transformed to career names, and returned with confidence percentages.

6. Frontend Page Architecture

Page	Route	Purpose
Landing	`/`	Hero section, feature highlights, CTA
Auth	`/auth`	Login/Register form with toggle
Profile	`/profile`	Score sliders, skill/interest tag selection → submit
Dashboard	`/dashboard`	Top 3 career prediction cards with confidence bars
Roadmap	`/roadmap/:career`	Step-by-step animated career roadmap timeline

7. Testing Strategy

Test Type	Method	Scope
API Smoke Test	`curl http://localhost:5000/api/health`	Backend is running, model trained
Registration Test	`curl POST /api/register` with JSON body	User creation + JWT return
Prediction Test	`curl POST /api/predict` with sample data	ML pipeline end-to-end
Full Flow (Manual)	Browser walk-through: register → profile → predict → roadmap	Complete user journey

8. Future Enhancements

Enhancement	Description
Real Dataset	Replace synthetic data with real student-career outcome data
Database	Add SQLite / PostgreSQL for persistent user storage
Model Persistence	Save/load trained models with `joblib` instead of retraining on every startup
More Careers	Expand from 12 to 20+ career domains
Feedback Loop	Let users rate predictions to improve model accuracy over time
Deployment	Deploy backend on Render/Railway, frontend on Vercel/Netlify

Name		Name	Last commit message	Last commit date
Latest commit History 3 Commits
backend		backend
frontend		frontend
.gitignore		.gitignore
README.md		README.md

Folders and files

Latest commit

History

Repository files navigation

🎯 CareerPath AI — Career Guidance System

📸 Screenshots

Landing Page

Profile Assessment

AI Career Predictions

Personalized Roadmap

🏗️ System Architecture

🔄 Workflow

🛠️ Tech Stack

Backend

Frontend

📁 Project Structure

🚀 Getting Started

Prerequisites

1. Clone the Repository

2. Setup Backend

3. Setup Frontend

4. Open the App

📡 API Endpoints

Example: Predict Career

🤖 ML Model Details

Ensemble Architecture

Training Data

Career Domains

🗺️ Personalized Roadmaps

✨ Features

🧪 Testing

Backend API Test

Full Flow Test

📄 License

👤 Author

📋 Project Methodology

1. Introduction & Problem Statement

2. Development Methodology

Phase 1 — Requirement Analysis & Data Modelling

Phase 2 — Machine Learning Model Development

Phase 3 — Backend API Development (Flask)

Phase 4 — Frontend Development (React + Vite)

Phase 5 — Integration & Testing

3. Technology Stack Justification

4. Data Flow — Complete User Journey

5. ML Model Methodology

5.1 Synthetic Data Generation

5.2 Model Training

5.3 Ensemble Prediction

6. Frontend Page Architecture

7. Testing Strategy

8. Future Enhancements

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Packages