Using DBT & DuckDB to manage Microsoft Fabric transformation layer
Join Fabric Dojo, become a Fabric Data Engineer: https://skool.com/fabricdojo/about
Want to learn Fabric AI Workflows? Join here: https://skool.com/fabricai/about
Follow along as I build a Microsoft Fabric data platform for by business (from scratch), exploring some of the latest Fabric & AI tools & processes, like Claude Code, BMAD, DuckDB, dbt, Ontologies, Data Agents & more.
Warning: this is an experimental project where I'm testing the latest tools and techniques and applying them to a real Fabric project build. I don't recommend you apply these techniques for your production workloads.
Mim's blog post that I used in the research: https://datamonkeysite.com/2026/03/05...
My own repo which the changes I made: https://github.com/UnifiedEducation/r...
Chapters
0:00 Introduction
2:40 dbt Foundations — what is dbt?
7:40 dbt integration uptions in Fabric
9:20 The DuckDB adapter explained
10:50 Feasibility research & Mim's work
17:40 Understanding Mim's example repo — Project Structure
19:00 Delta table generation with DuckLake + delta_export()
20:09 metadata.db & dynamic paths
21:40 dbt profiles.yml — dev, CI & Fabric (Mim's example)
22:40 In-memory engine & scale considerations
23:20 My POC — custom requirements & changes
25:50 Deployment workflow overview
29:00 Fabric notebook to run dbt
30:30 GitHub Actions for CI/CD (deploy-to-dev and deploy-to-prod)
33:12 Claude Skill example development workflow
44:10 Closing thoughts