This end-to-end MLOps project demonstrates production-grade machine learning infrastructure for weather forecasting. The pipeline ingests meteorological data from multiple sources, processes it through automated ETL workflows orchestrated by Apache Airflow, and trains prediction models with full experiment tracking. DVC (Data Version Control) manages dataset versioning and model artifacts, ensuring reproducibility across all experiments. The trained models are containerized and deployed to a Kubernetes cluster with auto-scaling capabilities to handle varying prediction loads. The frontend application, hosted on Vercel, provides users with intuitive weather forecasts, historical trends, and accuracy metrics. CI/CD pipelines automate testing, model validation, and deployment, while monitoring dashboards track model drift and performance degradation over time. This project serves as a reference implementation for scalable ML systems in production.