MLOps Production Pipeline

End-to-end MLOps production pipeline with experiment tracking, model registry, REST API serving, monitoring dashboard, and containerized deployment.

---

MLOps Lifecycle

Data Ingestion → Preprocessing → Experiment Tracking (MLflow) → Model Registry → API Serving → Monitoring & Drift Detection
       ↑                                                                                                     │
       └───────────────────────────────────── Retrain ───────────────────────────────────────────────────────┘

Features

• Experiment Tracking — MLflow auto-logging with metrics (accuracy, precision, recall, F1, ROC-AUC)
• Model Registry — Versioned model storage with stage transitions (Staging → Production)
• Hyperparameter Tuning — GridSearchCV with configurable param grids
• REST API Serving — FastAPI with /predict, /predict-batch, /model-info, /health, /metrics
• Monitoring Dashboard — Streamlit UI with performance trends, drift alerts, retrain trigger
• Data Drift Detection — KS test & Population Stability Index (PSI) for feature distribution shifts
• Prometheus Metrics — Prediction count, latency, error counters with Grafana dashboards
• Docker Deployment — Multi-service orchestration (MLflow, API, Monitoring, Prometheus, Grafana)
• CI/CD Ready — Test suite included, containerized, version-controlled

Architecture

┌─────────────┐     ┌──────────────┐     ┌──────────────┐     ┌──────────────┐
│   Training   │────▶│   MLflow     │────▶│  FastAPI     │────▶│  Prometheus  │
│  Container   │     │   Server     │     │  Serving     │     │  & Grafana   │
└─────────────┘     └──────────────┘     └──────────────┘     └──────────────┘
                           │                      │
                           ▼                      ▼
                    ┌──────────────┐     ┌──────────────┐
                    │   Model      │     │  Streamlit   │
                    │   Registry   │     │  Dashboard   │
                    └──────────────┘     └──────────────┘

Quick Start (Docker Compose)

# Clone the repository
git clone https://github.com/roohan-514/mlops-production-pipeline.git
cd mlops-production-pipeline

# Start all services
docker-compose up -d

# Access services:
# - MLflow UI:     http://localhost:5000
# - API:           http://localhost:8000
# - Dashboard:     http://localhost:8501
# - Prometheus:    http://localhost:9090
# - Grafana:       http://localhost:3000 (admin/admin)

Manual Setup

# Create virtual environment
python -m venv venv
source venv/bin/activate  # Linux/Mac
# venv\Scripts\activate   # Windows

# Install dependencies
pip install -r requirements.txt

# Start MLflow server
mlflow server --host 0.0.0.0 --port 5000 --backend-store-uri sqlite:///mlflow.db

# Train a model (in another terminal)
python -m pipeline.train

# Start the API
uvicorn api.main:app --host 0.0.0.0 --port 8000

# Start monitoring dashboard
streamlit run monitoring/dashboard.py --server.port 8501

Training a New Model

python -m pipeline.train

Configure training parameters in pipeline/config.py. The script:

1. Loads and preprocesses the dataset (Breast Cancer by default)
2. Runs GridSearchCV with Random Forest
3. Logs parameters, metrics, and model artifacts to MLflow
4. Registers the best model in the MLflow Model Registry

Running the API

uvicorn api.main:app --host 0.0.0.0 --port 8000 --reload

API Endpoints

Method	Endpoint	Description
POST	/predict	Single prediction
POST	/predict-batch	Batch predictions
GET	/model-info	Model metadata
GET	/health	Health check
GET	/metrics	Prometheus metrics

Example API Calls

# Single prediction
curl -X POST "http://localhost:8000/predict" \
  -H "Content-Type: application/json" \
  -d '{"features": [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0, 17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0, 25.0, 26.0, 27.0, 28.0, 29.0, 30.0]}'

# Batch prediction
curl -X POST "http://localhost:8000/predict-batch" \
  -H "Content-Type: application/json" \
  -d '{"instances": [[1.0, ...], [2.0, ...]]}'

# Health check
curl "http://localhost:8000/health"

Interactive API Docs

Open http://localhost:8000/docs in your browser for Swagger UI.

Monitoring Dashboard

The Streamlit dashboard provides:

• Overview — Model version, run count, API status
• Model Performance — Metrics across MLflow runs
• Prediction Distribution — Class distribution and confidence scores
• Data Drift Detection — KS test and PSI analysis per feature
• Recent Predictions — Prediction log viewer
• Retrain Model — One-click retraining trigger

Project Structure

mlops-production-pipeline/
├── pipeline/             # Training pipeline
│   ├── train.py          # Model training with MLflow
│   ├── evaluate.py       # Model evaluation & comparison
│   ├── data_preprocessing.py  # Data loading & preprocessing
│   └── config.py         # Training configuration
├── api/                  # REST API serving
│   ├── main.py           # FastAPI application
│   ├── config.py         # API configuration
│   └── schemas.py        # Pydantic models
├── monitoring/           # Monitoring & observability
│   ├── dashboard.py      # Streamlit dashboard
│   ├── metrics.py        # Prometheus metrics
│   └── drift_detection.py  # Drift detection
├── data/                 # Datasets
│   ├── sample_data.py    # Data generation
│   └── README.md
├── tests/                # Test suite
│   ├── test_pipeline.py  # Pipeline tests
│   └── test_api.py       # API tests
├── notebooks/            # Jupyter notebooks
├── Dockerfile.api        # API container
├── Dockerfile.train      # Training container
├── docker-compose.yml    # Multi-service orchestration
├── prometheus.yml        # Prometheus config
├── requirements.txt      # Python dependencies
└── README.md

Tech Stack

Component	Technology
Experiment Tracking	MLflow
Model Training	Scikit-learn, Pandas
API Serving	FastAPI, Uvicorn
Monitoring UI	Streamlit
Observability	Prometheus, Grafana
Deployment	Docker, Docker Compose
Drift Detection	SciPy (KS test, PSI)
Data Processing	Scikit-learn, Pandas, NumPy

Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

License

MIT