DBT Foundations: How Production Data Models Are Structured

Hosted by: Prabhakaran Vijayanagulu

Bootcamp Exclusive

Sat, 18 Jul 2026
11:00 AM - 01:00 PM IST

Online (Zoom)

Hosted by

Prabhakaran Vijayanagulu

Sr. Data Engineer @Grab

Share With Friends

You have written plenty of SQL by now, but a folder of scripts is not a project. There is no clean way to reuse a query, test it, or trace which model feeds which dashboard. dbt is how you cross that line. Before you build with it, you need a clear picture of how a dbt project is put together and why it is shaped that way.

This is the first session in a three-part dbt series, and it lays the foundations. We will walk through how a dbt project is structured, how models are layered, and how Jinja and macros work, so the hands-on building in the next two sessions makes complete sense. Expect clear explanations and worked examples, with one simple macro to see the ideas in action.

No setup needed for this session, just come ready to follow along. We will get dbt running on your machine in the next session, before we start building.

What we'll cover

🔹 How a dbt project is structured: models, seeds, macros, tests, and docs, plus dbt_project.yml and profiles.yml

🔹 Materializations at a glance: view, table, incremental, and ephemeral, and when each one fits

🔹 Layered modelling from staging to intermediate to marts, and why ref() and source() keep your lineage clean

🔹 Jinja basics, variables, loops, and conditionals, then a simple macro to see it work

🔹 A first look at testing and docs, plus resources to go deeper before the next session

Pre-requisites

✔️ Comfortable writing SQL: joins, aggregations, and CTEs

✔️ A basic understanding of how a data warehouse holds tables and views

✔️ No dbt experience needed, this session starts from the foundations