Technologies¶
Last reviewed: October 2026
Deep-dive pages on the data and serving platforms that come up most in senior data, platform and AI engineering interviews. Each page covers the architecture, core patterns, performance and cost tuning, governance and interview questions. The Databricks, dbt and FastAPI pages also include production incidents and a 60-second ramp checklist for the night before.
Who it's for: engineers who list these tools on their CV and need to defend design choices at staff or principal depth, and engineers who are moving into a role built on one of these platforms.
How the pieces fit¶
flowchart LR
SRC[(Sources)] --> ING[Ingest / CDC]
ING --> SF[Snowflake]
ING --> DB[Databricks]
DBT[dbt: transform, test, document] --> SF
DBT --> DB
SF --> API[FastAPI: serve data + LLM endpoints]
DB --> API
API --> APP[Apps / BI / GenAI]
Suggested learning path¶
- Start with the warehouse you use or are interviewing for, either Snowflake or Databricks. Read the architecture section, then the performance and cost section.
- Add dbt, because modelling, testing and CI for analytics come up in nearly every data platform loop.
- Finish with FastAPI to cover how data and GenAI features are served to applications.
- Practise out loud with the matching interview Q&A pages linked below.
-
Snowflake
Storage and compute separation, CDC with Streams and Tasks, Time Travel, Cortex AI, cost and performance tuning, and governance.
-
Databricks
Lakehouse on Delta Lake, the Spark execution model, Unity Catalog, Mosaic AI and Genie, and tuning.
-
dbt
Models and materializations, tests and contracts, snapshots and macros, and CI/CD for analytics engineering.
-
FastAPI
The async ASGI model, Pydantic validation, dependency injection and production-shaped LLM/RAG endpoints.
Go deeper¶
- Snowflake Cortex: AI SQL functions, Cortex Analyst and Search, and agents inside Snowflake.
- Interview Q&A: Snowflake · Databricks · dbt · SQL · Python · Data Engineering
- Role path: Data Platform.