---

🚗 Vehicle Insurance Claim Prediction MLOps Platform

An end-to-end Production-Ready MLOps Pipeline built using Python, MongoDB, AWS, Docker, GitHub Actions, CI/CD, and Machine Learning that automates the complete lifecycle from data ingestion to cloud deployment.

---

📌 Project Overview

Traditional Machine Learning projects usually end after training a model.

This project goes several steps further by implementing a complete production-grade MLOps architecture capable of:

• Automated Data Ingestion
• Data Validation
• Feature Engineering
• Model Training
• Model Evaluation
• Model Registry
• Model Versioning
• Cloud Storage
• Prediction Pipeline
• Docker Containerization
• Continuous Integration
• Continuous Deployment
• AWS Cloud Deployment

The entire workflow is designed following modular software engineering practices, making the project scalable, maintainable, and production-ready.

---

🏗 Complete Architecture

MongoDB Atlas
       │
       ▼
Data Ingestion
       │
       ▼
Data Validation
       │
       ▼
Data Transformation
       │
       ▼
Model Training
       │
       ▼
Model Evaluation
       │
       ▼
Model Registry (AWS S3)
       │
       ▼
Prediction Pipeline
       │
       ▼
Flask Web App
       │
       ▼
Docker Container
       │
       ▼
GitHub Actions
       │
       ▼
AWS EC2 Deployment

---

✨ Key Features

📥 Data Ingestion

• Fetches dataset directly from MongoDB Atlas
• Converts MongoDB documents into Pandas DataFrame
• Creates train-test split automatically
• Stores artifacts for downstream pipeline

---

✅ Data Validation

• Schema Validation
• Missing Value Validation
• Data Drift Detection
• Column Validation
• Numerical Feature Validation
• Categorical Feature Validation

---

🔄 Data Transformation

• Feature Engineering
• Missing Value Imputation
• Feature Scaling
• Encoding
• Pipeline Serialization
• Transformation Object Saving

---

🤖 Model Training

Multiple algorithms can be trained and compared.

Supports:

• Random Forest
• Gradient Boosting
• XGBoost
• CatBoost
• Extra Trees
• Decision Tree
• AdaBoost
• Linear Models

Automatic Best Model Selection based on evaluation metrics.

---

📊 Model Evaluation

Instead of replacing models blindly,

the project:

• Downloads previous production model
• Compares new model performance
• Uses Threshold Based Validation
• Registers model only if performance improves

This mimics a real production ML workflow.

---

☁ AWS Model Registry

Models are automatically:

• Uploaded to AWS S3
• Version Controlled
• Retrieved during prediction
• Managed without manual intervention

---

🔮 Prediction Pipeline

Production prediction pipeline supports:

• Single Prediction
• Batch Prediction
• Real-time Inference

---

🌐 Flask Web Application

Interactive web interface allowing users to

• Upload input
• Predict Insurance Claim
• Trigger Training Pipeline
• View Prediction Results

---

🐳 Dockerized Deployment

Entire application is containerized using Docker ensuring

• Environment consistency
• Easy deployment
• Platform independence

---

⚙ CI/CD Pipeline

Fully automated deployment using

GitHub Actions

Workflow:

Code Push
     │
     ▼
GitHub Actions
     │
     ▼
Build Docker Image
     │
     ▼
Push Image to AWS ECR
     │
     ▼
Deploy on EC2
     │
     ▼
Updated Application

No manual deployment required.

---

🛠 Tech Stack

Programming

• Python
• HTML
• CSS

---

Machine Learning

• Scikit-Learn
• Pandas
• NumPy

---

Database

• MongoDB Atlas

---

Cloud

• AWS EC2
• AWS S3
• AWS ECR
• AWS IAM

---

MLOps

• Docker
• GitHub Actions
• CI/CD
• Model Registry
• Environment Variables
• Logging
• Exception Handling

---

Backend

• Flask

---

Version Control

• Git
• GitHub

---

📂 Project Structure

Vehicle-Insurance-Project/

│
├── artifacts/
├── notebooks/
├── src/
│
├── components/
│     ├── Data Ingestion
│     ├── Data Validation
│     ├── Data Transformation
│     ├── Model Trainer
│     ├── Model Evaluation
│     ├── Model Pusher
│
├── configuration/
├── constants/
├── entity/
├── pipeline/
├── aws_storage/
├── utils/
│
├── static/
├── templates/
│
├── app.py
├── demo.py
├── requirements.txt
├── Dockerfile
├── setup.py
├── pyproject.toml
└── README.md

---

⚙ Pipeline Workflow

MongoDB
   │
   ▼
Data Ingestion
   │
   ▼
Data Validation
   │
   ▼
Data Transformation
   │
   ▼
Model Trainer
   │
   ▼
Model Evaluation
   │
   ▼
AWS S3 Model Registry
   │
   ▼
Prediction Pipeline

---

🔐 Environment Variables

Required environment variables

MONGODB_URL

AWS_ACCESS_KEY_ID

AWS_SECRET_ACCESS_KEY

AWS_DEFAULT_REGION

ECR_REPO

---

☁ AWS Services Used

• IAM
• EC2
• S3
• ECR

---

🐳 Deployment Pipeline

Developer

↓

GitHub Repository

↓

GitHub Actions

↓

Docker Build

↓

Push Docker Image

↓

AWS ECR

↓

EC2 Pulls Latest Image

↓

Application Deployment

---

🚀 How to Run

Clone Repository

git clone <repo_url>

---

Create Environment

conda create -n vehicle python=3.10

conda activate vehicle

---

Install Dependencies

pip install -r requirements.txt

---

Run Training

python demo.py

---

Run Application

python app.py

---

📈 Engineering Highlights

✔ Modular Pipeline Architecture

✔ Production-ready Folder Structure

✔ Config Driven Development

✔ Artifact Based Pipeline

✔ Object-Oriented Design

✔ Cloud Model Registry

✔ Automatic Model Comparison

✔ Automated Deployment

✔ Logging System

✔ Custom Exception Handling

✔ Dockerized Infrastructure

✔ Continuous Integration

✔ Continuous Deployment

✔ AWS Cloud Deployment

✔ Environment Variable Management

---

📚 Software Engineering Concepts Demonstrated

• Clean Code Principles
• Modular Architecture
• Separation of Concerns
• Dependency Management
• Configuration Management
• Pipeline Design Pattern
• Artifact Management
• Model Versioning
• Cloud Integration
• CI/CD Automation
• Containerization
• Production Deployment

---

🎯 Recruiter Highlights

This project demonstrates hands-on experience with:

Category	Skills Demonstrated
Machine Learning	End-to-End ML Pipeline
MLOps	Production Workflow
Backend	Flask
Cloud	AWS (EC2, S3, ECR, IAM)
Database	MongoDB Atlas
DevOps	Docker, GitHub Actions
CI/CD	Automated Deployment
Software Engineering	OOP, Modular Design, Logging, Exception Handling

---

👨‍💻 Author

Ayush

Aspiring Machine Learning Engineer | MLOps Engineer

Focused on building production-grade AI systems using Machine Learning, Cloud Computing, and DevOps practices.

---