Databricks Lakehouse Architecture
training code: DBX-LKH / ENG DL 1d / EN
The Databricks Lakehouse training is the third step in the structured training path Fundamental → Explorer → Lakehouse → Transformation. Participants will learn how to design and implement Lakehouse architecture using Delta Lake, Auto Loader, and Structured Streaming. The training also covers cost management, optimization, and governance.
The training is intended for data engineers and DataOps teams who want to learn how to build and maintain Lakehouse architecture and data processing workflows in Databricks.
– understand the Lakehouse concept and Medallion Architecture
– can create, load, and update data in Delta Lake
– are familiar with batch and streaming data ingestion techniques
– can optimize and monitor Delta tables
– understand governance and lineage in Unity Catalog
– know how to combine transformations, optimization, and quality control in practical workflows
– are prepared for the next stage of the training path – Databricks Transformation
1.Introduction to Lakehouse architecture
• Lakehouse concept – combining Data Lake and Data Warehouse
• Medallion Architecture structure (Bronze, Silver, Gold)
• The role of Delta Lake and Unity Catalog in data management
• Designing data flow logic between layers
2.Delta Lake in practice
• ACID operations and schema enforcement
• MERGE, UPDATE, DELETE, and INSERT – modifying Delta tables
• Time travel and change history (DESCRIBE HISTORY)
• Creating managed and external tables in Unity Catalog
3.Data ingestion – batch and stream
• COPY INTO as a batch data loading method
• Auto Loader (cloudFiles) – incremental ingest and schema evolution
• Monitoring streams in the new Streaming UI
4.Optimization and data management
• OPTIMIZE, ZORDER, and VACUUM – Delta Lake optimization mechanisms
• Partitioning and query plan analysis
• Liquid Clustering – automatic data clustering
• Delta Sharing – sharing data across teams and environments
5.Cost management (practical)
• Batch vs streaming costs (Auto Loader, Structured Streaming)
• Impact of OPTIMIZE, ZORDER, and VACUUM on costs
• Resource planning for large tables and pipelines
• Cost architecture in the Bronze–Silver–Gold model
6.Observability & Monitoring (light)
• Monitoring in Streaming UI and Metrics UI
• Alerts in Workflows and SQL dashboards
• Best practices for observability in Lakehouse
7.Security fundamentals (light)
• Row-level security and column masking – basics
• Token passthrough – awareness and scenarios
• Unity Catalog as a governance layer
8.Final project
• Design and implement a mini-Lakehouse with batch and stream data loading, Delta optimization, quality control, monitoring, and cost management
– Completion of Databricks Explorer or equivalent knowledge
– Experience with SQL and basic PySpark
– Basic understanding of cloud and data architecture concepts
-
access to Altkom Akademia student
Training method:
The training is conducted in the Databricks cloud environment. Each participant receives their own workspace with access to Unity Catalog, SQL Editor, Notebooks, and a catalog with test data.
Training: English
-
Materials: English