Track

Data Architecture Design

How the pieces of a data platform fit together and where each one actually runs. Every design starts with a full-size diagram, then explains the run flow, each component's job, the configuration files and how to describe the design in an interview.

3 chapters ~1.5 hours Snowflake · dbt · Airflow · AWS · Linux Level: intermediate → advanced

Designs

Design · Modern ELT

dbt + Airflow + Snowflake Architecture

The logical ELT reference architecture, then two production deployments: Airflow on Kubernetes (EKS) with one container image, and Airflow on Linux VMs with systemd, Celery and Ansible.

How to read the diagrams

Colour / mark	Meaning
AFTeal	Airflow: orchestration and its runtime
dbtOrange	dbt and the build/release path (code, CI/CD, Ansible)
SFBlue	Snowflake: storage, compute and access control
—Grey arrows	Data flow or a network connection
Numbered circles	Match the numbered steps under each diagram
Dashed outline	A network boundary (VPC, private network)

Use ⤢ Full screen on any diagram to see it at full size. On a phone, scroll the diagram sideways.