FlowForge

an MLOps Project

A production-ready MLOps pipeline demonstrating best practices for machine learning operations, from training to deployment and monitoring.

Overview

This project implements a complete MLOps workflow for a sentiment analysis model, showcasing:

• Automated CI/CD with GitHub Actions
• Model Registry with MLflow
• Containerized Training with Docker
• REST API for inference (FastAPI)
• Web Interface for interactive demos
• Model Monitoring and drift detection

Project Structure

mlops/
├── src/
│   ├── data/          # Data processing and loading utilities
│   ├── models/         # Model definitions and architectures
│   ├── training/       # Training scripts and pipelines
│   └── inference/      # API and inference code
├── tests/              # Unit and integration tests
├── docs/               # Documentation and architecture notes
├── scripts/            # Utility scripts (health checks, etc.)
├── .github/            # GitHub Actions workflows
└── requirements.txt    # Python dependencies

Quick Start

Verify Setup (Health Check)

Before running anything, verify your setup is correct:

Using Docker:

# Build the image
docker build -t mlops:latest .

# Run health check
docker run --rm mlops:latest python scripts/health_check.py

Or use the test script:

chmod +x scripts/test_docker.sh
./scripts/test_docker.sh

Local (if you have Python installed):

python scripts/health_check.py

Using Docker (Recommended)

Since this project is containerized, you can run everything without installing Python locally:

1. Build the Docker image:

   docker build -t mlops:latest .

2. Run training:

   docker run --rm mlops:latest python src/training/train.py

3. Run the inference API:

   docker run -p 8000:8000 mlops:latest python src/inference/app.py

Local Development (Optional)

If you have Python 3.10+ installed locally:

1. Create a virtual environment:

   python3.10 -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate

2. Install dependencies:

   pip install -r requirements.txt

3. Run training:

   python src/training/train.py

4. Start the inference API:

   python src/inference/app.py

Documentation

See the /docs directory for detailed documentation:

• Architecture Overview
• MLOps Stages
• Data Versioning
• Design Decisions

License

MIT