Databricks Workshops

Description of some databricks workshops and learning material we have developed at Knowit.

Workshops (Knowit toppturer)

These workshops are 2.5h hands-on workshops for learning various important aspects of databricks.

At Knowit we call these workshops Toppturer, giving quick but meaningful experience with a technology/tool/framework.

Available workshops:

• Workshop: Data engineering on Databricks
• Workshop: Using LangChain and open LLM-models on Databricks
• Workshop: LLM Adaptation on Databricks
• Workshop: DataOps on Databricks, using git and versioning of tables, jobs and code

Workshop: Data engineering on Databricks

Link: https://github.com/knowit/AWS-Databricks-NYC-Taxi-Workshop

For: Developers, analysts, data scientists, data engineers.

Pre-requisites: Some python knowledge

Topics:

• Basic understanding of components and tools in Databricks
• Perform data transformation in Spark SQL and Pyspark
• Use Databricks Reops for git-versioned Data Engineering
• Deploy a Spark job with Databricks Workflows
• Write ETL code and data quality checks in Delta Live Tables

Link:

Workshop: Using LangChain and open LLM-models on Databricks

Link: https://github.com/paalvibe/llm-langchain-course

For: Anybody

Topics:

• Setup and use of LLMs in Databricks
• Use of Langchain-rammeverket for:
  • LLM-wrapping
  • LLM-serving
  • Summarizing
• Context embedding with chromadb
• Reformating
• Multi query retrieval
• Prompt engineering

Workshop: LLM Adaptation on Databricks

Link: https://github.com/paalvibe/llm-tune-course

For: Anybody

Topics:

• What is an LLM (Large Language Model)?
• Tuning of LLM models on Databricks
• Different modes of adapting LLMs
• When and when not to train your own LLM?

Workshop: DataOps on Databricks, using git and versioning of tables, jobs and code

Link: https://github.com/paalvibe/databricks-dataops-course

For: Data Engineers, Full stack data scientists, ML Engineers, Data Platform Engineers

Topics:

• Opinionated git-based approach to DataOps
• Structure your environments to allow for dev runs of data pipelines
• Move data pipelines from dev to prod
• Using git branches and commits to name and manage data and jobs responsibly
• Will not do Github Actions here, but the processed needed are used
• Does not cover data quality nor pipeline management

Pre-requisites: Some python knowledge

FUTURE Workshop: DataOps on Databricks part 2

For: Data Engineers, Full stack data scientists, ML Engineers, Data Platform Engineers

• How to enable data contracts and data quality checks in pipelines
• Difference between Delta Live Tables and regular databricks notebooks

Pre-requisites: Some python knowledge