The analytics engineer is the newest of these titles. dbt 37,942 Labs, then Fishtown Analytics, popularized it in a March 2019 blog post by Claire Carroll, which described someone who transforms, tests, deploys and documents data and applies "software engineering best practices like version control and continuous integration to the analytics code base". The post placed the role "somewhere in the middle" between data engineering and analysis. Cheap cloud warehouses and low-code ingestion tools had left one job open: turning raw tables into business models in SQL.
The working split today is roughly this: the data engineer owns everything up to the raw tables in the warehouse or lake (ingestion, storage, streaming, orchestration, performance and cost), and the analytics engineer owns the SQL models on top (cleaned staging tables, facts and dimensions, metric definitions and their tests). At BookNest, a data engineer makes sure every order lands in the warehouse within minutes; an analytics engineer decides whether a refunded order counts as revenue and encodes that rule once, in a tested model, so every dashboard agrees.
The tool that defines the role is dbt, which compiles SELECT statements into tables and views and runs tests against them. Analytical SQL and Data Warehouses builds BookNest's sales mart with it (Transforming Data with dbt), and Orchestration and Pipelines schedules it (Orchestrating dbt with Cosmos).