Summary
What you’ll impact
Our organization is seeking a Staff Data Architect to own and enhance the semantic layer of its data platform, ensuring clean, AI-ready models that serve as the single source of truth. The role involves designing metrics, driving data quality, enabling self-serve analytics, and collaborating across product, clinical, and business teams, all in a fully remote environment.
Responsibilities
What you'll do
- Own the semantic layer: Design and maintain the metrics, dimensions, and data models that serve as the single source of truth across the business — built to be queried by humans and AI alike.
- Architect for AI-readiness: Structure our Databricks and GCP-based data platform so it can reliably power AI-driven analytics agents — clean schemas in Unity Catalog, consistent naming conventions, and strong data contracts.
- Drive data quality end-to-end: Build observability, testing, and alerting into pipelines so we catch problems before stakeholders do.
- Enable self-serve analytics: Partner with product, clinical, and business teams to deliver tooling and dashboards that let them answer their own questions without filing tickets.
- Set the standard: Establish and enforce data modeling conventions, documentation practices, and code review norms across the data org.
- Collaborate cross-functionally: Translate ambiguous business questions into data solutions; work with product engineering to improve event logging and measurement coverage.
Requirements
What you’ll bring
- 8+ years of analytics engineering, data engineering, or data architecture experience, with clear progression to staff-level scope and impact.
- Deep SQL expertise and strong Python skills.
- Strong data modeling fundamentals — you know what clean, extensible, well-documented models look like and can build and enforce that standard regardless of tooling. Experience with dbt is a plus.
- Experience with modern data stack tools: Databricks, Snowflake or BigQuery, Airflow or similar orchestration, Looker/Tableau or equivalent BI tools.
- Track record of building resilient, monitored, production-grade data systems — not just pipelines that work, but ones that stay working.
- Strong communicator who can make technical tradeoffs legible to non-technical stakeholders.