Skip to content
Here Be Dragons

Home-health routing system, post series

This series follows the architecture behind a home-health scheduling and routing system. The through-line is practical: every technical choice is constrained by clinical data, regulated health information, agency tenancy, clinician skills, visit windows, and an operations UI that needs route updates without polling.

Reading order

  1. OASIS in home-health software, the assessment data model that shapes the product.
  2. HIPAA for software engineers, the privacy and security rules around the data.
  3. Multi-tenant Django that fails closed, row-level tenancy as the default query path.
  4. The state machine as a DRF action, visit workflow rules exposed through safe API transitions.
  5. What I learned from Google’s Vehicle Routing Problem docs, the optimization problem behind scheduling routes.
  6. OR-Tools VRP with skill constraints and time windows, the solver model for clinicians and visits.
  7. A scikit-learn re-ranker inside an OR-Tools objective, soft preference scoring inside a hard routing model.
  8. Redis three ways, background work, cache state, and WebSocket fanout.
Why the name "Here Be Dragons"?