Job Description
ABOUT THE ROLE
\n
The Enterprise Data Lake team is looking for a Data Modeler to turn approved business context into governed ontology. You take what the business analyst hands off (source discovery, business requirements, a modeler handoff addendum) and build the shared model that engineering builds from: business concepts, natural keys, relationships, and processes, defined clearly enough that a stakeholder recognizes their own business in it and an engineer can build from it without a walkthrough.
\n
We work as consultants embedded with a client team, so you will be largely self-directed and trusted to make the people around you more effective.
\n
WHAT YOU'LL OWN
\n
Our delivery follows a defined framework with clear ownership at each step. The deliverable that is fully yours is the Business Ontology. You also carry the model forward into the conceptual and physical layers that hand off to engineering.
\n
– Business Ontology. Yours to own, the business concepts, relationships, and processes that everyone downstream builds from.
\n
– Conceptual and Physical Models. Carry the approved ontology forward into an implementation-ready spec engineering can build from directly.
\n
– RDM Candidates. The formal candidates table required before architect approval.
\n
– Source Discovery and Requirements. A consumer, not an author. These come from the business analyst. If a handoff is incomplete, you push back for more context rather than doing the discovery work yourself.
\n
CORE RESPONSIBILITIES
\n
– Author the business ontology. Sequence concepts starting from a base case explainable in plain terms, distinguish standalone entities from attributes masquerading as entities, and keep concepts (nouns) and processes (verbs) clearly separated.
\n
– Build ER diagrams. Diagram each concept with natural keys marked, approved simplified data types only, correct cardinality notation, and a non-exhaustive attribute list.
\n
– Design and validate natural keys. Hypothesize natural keys from real-world, business-recognizable attributes, never system-generated IDs or classification-only fields, then test candidates against live data using cardinality and edge-case analysis, quantifying conflicts rather than just flagging them.
\n
– Apply data typing discipline. Distinguish free-text fields from category fields with a finite, governable set of values, and route every category field into the RDM candidates table.
\n
– Document relationships and subtypes. Name relationship type and cardinality in prose, and give any subtype with its own natural key or distinct attributes its own entity in the diagram rather than leaving it prose-only.
\n
– Hand off to engineering. Translate the approved ontology into source-to-target mappings and a physical model, so the concept moves cleanly into pipeline build.
\n
WHAT MAKES SOMEONE SUCCESSFUL HERE
\n
– Ontological thinking. You reach for the real-world business concept before the table, and you can explain why a natural key or a data type choice matters to someone without a technical background.
\n
– Comfort validating against real data. You do not treat a natural key hypothesis as settled until you have tested it, and you can read what live data is telling you about an edge case.
\n
– Structural rigor. Ontology documents that pass standards review without a rework loop over missing diagrams, mistyped fields, or undocumented subtypes.
\n
– Clarity on where the work hands off. A good sense of where source discovery ends and modeling begins, and the judgment to push back on an incomplete handoff instead of quietly filling the gap yourself.
\n
– Ability to ramp with limited context. You can start mid-flight in an unfamiliar domain and build a workable model without needing every fact spelled out.
\n
NICE TO HAVE
\n
Hands-on SQL and comfort profiling data in Databricks. Experience with ER diagramming tools such as Eraser.io. Exposure to graph, hierarchy, or recursive modeling. Familiarity with utility, billing, account, or customer-service data is a welcome addition rather than a requirement.
\n
WAYS OF WORKING AND GROWTH
\n
This is a full-time, team-embedded role: expect daily standups and modeling syncs with camera on, and regular pairing with the rest of the modeling team rather than solo, low-contact work. The seat sits in our transformation practice, with room to grow toward ownership of full domains end to end and, over time, mentoring the next modeler who joins.
\n
WHAT SUCCESS LOOKS LIKE EARLY ON
\n
– Your first ontology document clears standards review with only minor revisions.
\n
– A natural key you proposed holds up once tested against real data, or you catch and explain the edge case that breaks it.
\n
– You are comfortable pushing back on an incomplete handoff instead of guessing.
\n
– You are working comfortably within the delivery framework and partnering well with the business analyst and engineering team.
\n
\n
**Hybrid in Columbus, Ohio or Cincinnati, Ohio or Charlotte, North Carolina