$ cat projects/hims-agroha.mdx
HIMS - Healthcare Management System
Full-stack hospital and medical-college workflows: Haryana pay-matrix payroll, IPD operations, approval systems, laboratory integration, reporting, and faster data retrieval.
React · JavaScript · Redux Thunk · Node.js · Express · MongoDB · Redis · EJS · 2022 · live
Daffodil Software · May 2022 - Feb 2023
HIMS is a hospital and medical-college management platform used by clinical, administrative, finance, laboratory, pharmacy, and academic teams. I owned end-to-end delivery across React screens, Node.js/Express APIs, MongoDB data models, workflow logic, and server-generated reports for several modules.
what I built
- Haryana government pay-matrix payroll - redeveloped payroll logic and supporting master data for pay matrices, grade pay, configurable allowances, deductions, attendance-linked holds, arrears, monthly batch generation, salary slips, and finance reports. I worked directly with hospital stakeholders to turn requirements into a tested implementation.
- IPD operations - led an inpatient module covering admission, bed and room allocation, daily clinical records, medications, laboratory data, discharge, and death workflows. I guided two interns through implementation, debugging, and code review.
- Approval workflows - built role-based, auditable attendance and inventory approval flows with conditional rejection routing and downstream finance processing.
- About 20% faster patient and diagnostic retrieval - used MongoDB execution-plan analysis, compound indexes, simpler data-access queries, and Redis caching with a six-day TTL. This is an approximate before-and-after result.
- Laboratory and document workflows - integrated external laboratory submission and result polling, mapped results into an internal schema, and generated clinical, payroll, and operational PDFs from server-side EJS templates.
ownership boundaries
The platform predated my involvement. I extended the existing salary-generation pipeline and JWT middleware rather than creating them. I owned the Haryana pay-matrix logic, related APIs and screens, and module-level delivery; billing was outside the scope of my IPD ownership.
what stuck with me
- Healthcare workflows demand careful state transitions, resilient validation, and clear role boundaries because operational mistakes have real consequences.
- Good performance work starts with evidence: execution plans, data shape, and cache suitability before changing queries.
- Technical ownership is also making an unfamiliar module understandable enough for the next engineer to maintain.