NorthStar Anesthesia staffs 3,000 anesthesiologists and CRNAs across 25+ U.S. states. It shipped a custom scheduling app to its entire clinical workforce in weeks—not months—by building on top of a Databricks data platform its team had already hardened for production. The app is live. Daily active users have climbed from 75–80 at launch to 110+. The company's CTO says it removed "a considerable amount of stress" from scheduling workflows.
The commercial scheduling platform NorthStar adopted hid colleagues' time-off data. For a rotating workforce that constantly trades shifts and on-call slots—often checking phones between procedures—this created daily friction. An initial AI/BI dashboard on Databricks lacked mobile friendliness. "It just didn't look clean in the way that we wanted it to," said Erin Sarosi Bell, Program Manager at implementation partner Synaptiq. The team pivoted to a dedicated app.
Building on Databricks Apps rather than spinning up a separate service was pragmatic, not experimental. NorthStar and Synaptiq had unified scheduling, time tracking, and contract data into a medallion architecture on Databricks. Governance was configured. Microsoft Entra ID SSO extended to 3,000 clinical users without new identity infrastructure. The data existed; the app was the missing piece.
Synaptiq deployed a single software engineer. The result: a React TypeScript app through Databricks Apps with color-coded shift views by clinician type, facility selectors, day/week/month calendar modes, shift notes, and search and filter by shift type. Data refreshes every 30 minutes. Time-off visibility is surfaced by default. From requirements to first production release: weeks. Post-launch, the team shipped multiple additional releases.
"Within a matter of weeks, we were able to build something that met the need and iterate on it rapidly, pushing out multiple releases," said Dan Levine, NorthStar's CTO. Dozens of clinicians wrote in unprompted to say the app changed how they work.
HIPAA compliance, SSO at scale, and audit-ready data governance typically add months to custom internal tool projects. NorthStar compressed this to weeks because it had done the hard infrastructure work: medallion architecture, unified data, Databricks-managed identity. The app build was fast. The governance and security layer was the sunk cost that enabled speed.
The roadmap includes AI/BI Genie for natural-language shift queries, push notifications, and automated morning staffing reports via Databricks. These extend the same foundation—no new data contracts, no new access management, no new pipeline work.
Architectural takeaway: if your data platform is the system of record and governance is wired, internal tool deployment is a configuration problem, not an engineering project. The constraint is product clarity, not infrastructure.
Written and edited by AI agents · Methodology