The Analytics Engineer Role Is Still Misunderstood by Most Hiring Managers

Hot take: most companies hiring for 'analytics engineer' don't actually know what they're hiring for, and it shows in the job posting before the first interview even happens. I've reviewed dozens of these postings for clients trying to build out a data function, and they split cleanly into two camps, neither of which is actually the role.

The two miscast versions of the job

Camp one is a data analyst role — building dashboards, answering ad-hoc stakeholder questions, writing SQL against existing tables — relabeled 'analytics engineer' because the title sounds more technical and commands a better salary band in the current market. Camp two is a data engineering role — building and maintaining ingestion pipelines, managing warehouse infrastructure, owning Airflow DAGs — relabeled the same way because the hiring manager conflated 'engineer' with 'does engineering work,' full stop.

The analytics engineer role exists specifically in the gap between those two jobs. Mislabel either one with the title and you'll hire the wrong person for what you actually need.

What the role actually is

An analytics engineer's actual job is transforming raw ingested data into clean, tested, documented, reusable models that analysts and stakeholders build on top of — the dbt layer, specifically. They don't build the ingestion pipelines (that's data engineering) and they don't primarily build the dashboards (that's analytics). They own the modeling layer in between: defining what a 'customer' or an 'active session' means once, correctly, with tests that fail loudly when the definition breaks, so that every downstream dashboard inherits a consistent, trustworthy number instead of every analyst redefining it themselves.

• If the job posting's main responsibilities are 'build dashboards' and 'respond to stakeholder requests,' it's a data analyst role — hire for SQL and communication skills.
• If the main responsibilities are 'maintain our Airflow pipelines' and 'manage warehouse infrastructure,' it's a data engineering role — hire for systems and infra skills.
• If the role is actually 'own the dbt models, write data tests, define metrics once for the whole org,' that's the real analytics engineer job — hire for modeling discipline and a tester's instinct, not raw SQL speed.

I've watched companies hire a brilliant data analyst into a title they thought meant 'build us a semantic layer,' only to be disappointed when that person — who was never asked to do that in their career — didn't naturally produce one. The title didn't create the skill gap. The job posting did, three months before the interview even happened.