Modern enterprises require unified platforms that integrate data engineering, governance, and analytics to support strategic decision-making. This thesis evaluates Databricks as a cloud-native environment for implementing and extending a structured, metadata-driven ETL framework developed by Data Science Operations. The framework follows a modular three-layer architecture: raw data ingestion (L0), integration and quality control (L1), and analytical publication (L2). It was fully re-engineered on Databricks using Delta Lake for transactional consistency and Databricks Jobs for orchestration. Metadata-driven automation, dependency-aware scheduling, and comprehensive auditing mechanisms ensure traceability, reproducibility, and operational efficiency across the entire pipeline. Beyond traditional ETL operations, the study demonstrates integration with advanced analytics workflows.