AI Learning Path Recommender

An intelligent multi-service learning roadmap generator that creates personalized study plans using semantic search, vector embeddings, ChromaDB, and AI-powered topic matching.

The system recommends structured weekly learning roadmaps based on a user's:

• Topic / Subject
• Skill Level
• Weekly Study Hours
• Learning Duration

Instead of returning random courses, the system uses semantic similarity search and strict topic validation to generate more reliable and professional learning paths.

---

Project Architecture

This project follows a 3-tier architecture:

Frontend (React + Vite)
        ↓
ASP.NET Core Web API
        ↓
Python ML Service (FastAPI + ChromaDB)

---

Technologies Used

Frontend

• React.js
• Vite
• Axios
• CSS

Backend API

• ASP.NET Core Web API (.NET)

Machine Learning Service

• Python
• FastAPI
• SentenceTransformers
• ChromaDB
• Pandas
• NumPy
• Scikit-learn

AI / NLP

• Semantic Search
• Vector Embeddings
• Topic Resolution
• AI-assisted Subject Classification

---

Core Features

Personalized Learning Roadmaps

Generates weekly learning plans based on:

• User topic
• Skill level
• Weekly study hours
• Duration

---

Semantic Course Search

Uses SentenceTransformer embeddings to:

• Understand meaning instead of exact keywords
• Find relevant learning resources semantically

---

ChromaDB Vector Database

Stores vector embeddings of courses for:

• Fast retrieval
• Semantic similarity search
• Intelligent recommendations

---

Topic Validation System

Professional fallback handling:

• Unsupported topics are blocked
• No fake/random roadmap generation
• Users receive supported topic suggestions

Example:

{
  "status": "unsupported_topic",
  "message": "We do not have enough courses for this topic.",
  "available_topics": [
    "Machine Learning",
    "Web Development",
    "Data Science"
  ]
}

---

Available Topics Section

Frontend automatically loads supported roadmap topics from backend and displays them as clickable chips/tabs.

---

Multi-Service Communication

The application demonstrates real-world microservice communication:

• React → .NET API
• .NET API → Python ML Service

---

Project Structure

ai_assistant/
│
├── frontend/                # React + Vite Frontend
│
├── web_api/                 # ASP.NET Core API
│
├── ml_service/              # Python ML Service
│   ├── data/
│   ├── src/
│   ├── chroma_db/
│   └── requirements.txt
│
└── README.md

---

How the System Works

Step 1 — User Input

User enters:

• Topic
• Skill level
• Weekly hours
• Duration

Example:

Topic: Machine Learning
Level: Beginner
Hours/Week: 10
Duration: 6 Weeks

---

Step 2 — Frontend Sends Request

React frontend sends request to ASP.NET Core API.

---

Step 3 — .NET API Processing

The API:

• Validates request
• Forwards request to Python ML service

---

Step 4 — ML Service Processing

Python service:

• Resolves topic aliases
• Searches ChromaDB
• Retrieves semantically similar courses
• Builds structured roadmap

---

Step 5 — Response Returned

Frontend visualizes:

• Weekly roadmap
• Course sequence
• Estimated study hours
• Progress status

---

Semantic Search Flow

User Query
   ↓
SentenceTransformer Embedding
   ↓
ChromaDB Similarity Search
   ↓
Relevant Courses Retrieved
   ↓
Roadmap Generation

---

Machine Learning Concepts Used

Sentence Embeddings

Courses and user queries are converted into vectors using:

SentenceTransformer

This allows:

• Meaning-based retrieval
• Better recommendations
• NLP-powered matching

---

Vector Similarity Search

ChromaDB compares vector similarity to find the most relevant courses.

---

Topic Resolution

The system maps:

AI → Artificial Intelligence
ML → Machine Learning

and validates supported topics before generating a roadmap.

---

API Endpoints

Python ML Service

Health Check

GET /health

Available Topics

GET /topics

Generate Roadmap

POST /generate-roadmap

---

ASP.NET Core API

Health Check

GET /health

Available Topics

GET /api/learning-path/topics

Generate Learning Path

POST /api/learning-path/generate

---

Installation Guide

1. Clone Repository

git clone https://github.com/your-username/ai-learning-path-recommender.git
cd ai-learning-path-recommender

---

2. Setup Python ML Service

Navigate to ML Service

cd ml_service

Create Virtual Environment

Windows

python -m venv venv
venv\Scripts\activate

Mac/Linux

python3 -m venv venv
source venv/bin/activate

---

Install Dependencies

pip install -r requirements.txt

---

Run ML Service

uvicorn src.api.main:app --reload

Runs on:

http://127.0.0.1:8000

---

3. Setup ASP.NET Core API

Navigate to API

cd web_api

Restore Packages

dotnet restore

Run API

dotnet run --urls http://localhost:5000

Runs on:

http://localhost:5000

---

4. Setup Frontend

Navigate to Frontend

cd frontend

Install Dependencies

npm install

Run Frontend

npm run dev

OR

npm run build
npm run preview

Frontend URL:

http://127.0.0.1:5173

---

Running Full Application

Start services in this order:

Terminal 1 — ML Service

cd ml_service
uvicorn src.api.main:app --reload

---

Terminal 2 — ASP.NET API

cd web_api
dotnet run --urls http://localhost:5000

---

Terminal 3 — Frontend

cd frontend
npm run build
npm run preview

---

Example Workflow

User selects "Machine Learning"
        ↓
Frontend sends request
        ↓
.NET API receives request
        ↓
Python ML service processes query
        ↓
ChromaDB finds relevant courses
        ↓
Roadmap generated
        ↓
Frontend visualizes roadmap

---

Professional Design Decisions

Strict Unsupported Topic Handling

The system does NOT generate fake/random roadmaps.

If insufficient topic data exists:

• Roadmap generation stops
• User receives supported topic suggestions

This prevents low-quality recommendations.

---

Semantic Recommendation System

Instead of keyword matching:

• Uses embeddings
• Understands context and meaning
• Produces more relevant recommendations

---

Microservice Architecture

Separation of concerns:

• Frontend handles UI
• .NET handles API orchestration
• Python handles ML logic

This architecture is scalable and production-oriented.

---

Current Limitations

• Roadmap quality depends on dataset quality
• Some topics may have limited course coverage
• ChromaDB is currently local storage
• No authentication system yet

---

Future Improvements

• User authentication
• Progress tracking
• Roadmap export (PDF)
• AI chatbot mentor
• Course difficulty prediction
• Recommendation feedback loop
• Docker support
• Real-time analytics

---

Python Requirements (requirements.txt)

fastapi
uvicorn
pandas
numpy
scikit-learn
sentence-transformers
chromadb
pydantic
python-dotenv
anthropic
requests

---

Frontend Dependencies

npm install react react-dom axios
npm install -D vite

---

Learning Outcomes

This project demonstrates:

• NLP applications
• Semantic search systems
• Vector databases
• Recommendation systems
• Microservice communication
• API integration
• React frontend development
• ML service architecture

---

Author

Developed as an AI-powered educational roadmap generation system using:

• React
• ASP.NET Core
• FastAPI
• ChromaDB
• SentenceTransformers