Into backend and cloud systems: APIs, servers, infra.
A Go-based pipeline converting HL7v2 and FHIR R4 messages into a validated, queryable store: parsing, terminology validation against LOINC/SNOMED-CT, Postgres persistence, a GraphQL API, and a live SSE pipeline visualizer.
Built with Farrah, who handles QA while I work on the backend.
8-class classification pipeline on the Data Cortex Nuclear dataset, comparing Decision Tree, Random Forest, SVM, and XGBoost via stratified 5-fold CV, plus feature importance analysis.
A Django radiology information system. The front desk books a procedure, a technologist uploads the DICOM, and a radiologist reads it in an in-browser viewer built on Cornerstone.js and writes the report. Access is scoped to four staff roles.
A group project with Joseph and Deshny. We built the backend and the DICOM viewer; the rest of the frontend came from the client, a master's student.
Provisioned VMs and Samba AD/DC infrastructure on Proxmox, extended a federated GraphQL API (Go, gqlgen), designed Figma mockups for patient-facing modules.
Co-managed the org's budget and financial planning, and ran marketing campaigns for org events.