This repository implements a production‑grade end‑to‑end machine‑learning pipeline for time‑series forecasting using a custom dataset. It covers the full MLOps lifecycle:
- Data ingestion & feature engineering
- LSTM model training with transfer learning
- Experiment tracking with MLflow
- Asynchronous training jobs
- Real‑time inference API built with FastAPI (JWT auth, Redis caching, rate limiting)
- Monitoring via Prometheus and Grafana
- Docker Compose for reproducible environments
# Install dependencies with uv (recommended)
uv sync
# Copy example environment variables
cp .env
# Build and start services
docker compose up -d
# Run the API
uv run python -m src.api.mainThe application reads configuration from a .env file (see .env.example for a template). Required variables:
PROJECT_NAME=MLOps Pipeline
API_V1_STR=/api/v1
SECRET_KEY=your_secret_key_here
ACCESS_TOKEN_EXPIRE_MINUTES=30
REDIS_URL=redis://localhost:6379/0
MLFLOW_TRACKING_URI=http://localhost:5000
Adjust values as needed for your deployment.
- API docs:
http://localhost:8000/docs - MLflow UI:
http://localhost:5000 - Grafana:
http://localhost:3000(default credentials admin/admin)
- This project is licensed under the MIT License. See
LICENSEfor details.
