Pharmacy Transfer Pipeline

Overview

Onboarding a pharmacy partner required moving thousands of existing prescriptions into a new operating system. The prior process depended on a long sequence of manual steps, making large transfers difficult to monitor, pause, recover, or repeat safely.

What I Did

I designed and built the end-to-end transfer workflow, then evolved it from an initial import path into a production pipeline with administrative controls and an accompanying export capability.

• Decomposed each prescription import into independently retryable subtasks.
• Used SQS workers for asynchronous execution and idempotency controls to make retries safe.
• Added throttling, batch selection, pause and resume controls, and step-level progress reporting.
• Built file validation, conversion, and prescription-image rendering tools.
• Replaced operator scripts with a guided Django administration workflow.

Engineering Decisions

Large batch work rarely fails as one clean unit. Designing around resumable, idempotent steps made partial failure a routine operating condition instead of an exceptional recovery project.

Outcome

The pipeline imported more than 10,700 prescriptions across three partner onboardings. It replaced more than 14 manual steps with a workflow that operators could start, observe, pause, and resume without direct database or host access.