Analytics Engineer

The Position

OPSWAT's data platform turns business-system data into governed, documented, tested tables that business teams can read without asking what a column means. It is built on dbt Core, layered source → staging → intermediate → mart, and the warehouse is moving from Microsoft Fabric Warehouse to Azure Database for PostgreSQL. You will learn the craft on the real thing: models that ship, with value parity proven before anything is cut over.

This is the entry point to analytics engineering on this team, and it is a working role rather than an observation one. You will build models for the domains assigned to you and they will go into production. What differs from a more senior role is the supervision, not the bar: every change is reviewed before it lands, you have a named mentor, and you are paired with a Product Analyst on the business side: the two halves answer for the same number.

By day 60 you are expected to be delivering a model with less correction than your first one, not to still be waiting for instructions.

What You Will Be Doing

• Build staging, intermediate and mart dbt models for the domains assigned to you, following the team's layer standards: what belongs in each layer, where collation is normalised, one path in per source.
• Write tests that fail when the data breaks, not checks that pass whatever the data does.
• Run parity comparisons between a dbt model and the legacy procedure it replaces: execute the comparison, record the numbers, and learn what "proven" means before a consumer is repointed.
• Investigate the quality failures your tests surface, and trace them back to the source instead of patching the symptom in the model.
• Keep documentation and lineage current as models change, so the next reader does not have to ask what a column means.
• Work across both engines while the warehouse migrates, and notice when the same model behaves differently on each.
• Work in an AI-assisted engineering workflow. You set the task, review the output, and own what lands.
• Pair with a Product Analyst to learn what the business needs from a number, and why a definition is worth arguing about.
• Ask for review early, and act on it.

What we need from you

• 0–2 years in analytics engineering, data engineering, or a comparable role or a portfolio that shows the same work.
• Working SQL: joins, filters, aggregations, CTEs. Window functions are the next thing you learn, and you are already reading about them.
• Some Python for data work. A script you wrote, a notebook you finished, a pipeline you broke and then fixed.
• Git and pull-request basics, or the willingness to learn them in the first week.
• The habit of checking your own numbers before sharing them: reconciliation, sanity checks, and saying so when something does not add up.
• Curiosity about how the business actually works. You want to know why the data looks the way it does.
• Willingness to ask for help early. On this team that is the norm, not a weakness.

It Would Be Great If You Also Have

• Exposure to dbt, or another transformation framework, from coursework or a project.
• A BI tool you have built in: Superset, Power BI, Looker, or similar.
• A data project you can talk about for ten minutes: what it was for, what you got wrong, and what you changed.
• Comfort in a terminal: Docker, a CLI, a cloud console.
• Azure basics: storage, identity, or a service you have actually used

What success looks like in your first 60 days

• Your first models are merged, reviewed, and pass the layer-standards check. You learned the rules by applying them.
• A test you wrote catches something real, or would fail if the model broke.
• You have run a parity comparison end to end and can explain what "proven" means without reading the definition back.
• You can explain your domain's data in business terms, and why the definition matters.
• Your reviewer's comments are getting fewer and smaller: the next model needs less correction than the first.