Designed and maintained a Sanity-based content platform and TypeScript publishing pipeline that enabled teams to author, validate, localize, migrate, and publish structured digital-health content into Firebase-backed mobile applications.
Abstract architecture diagram — no product screenshots
Overview
The platform provided a structured authoring environment for a digital-health mobile application. Editors managed hierarchical program content across development, QA, staging, and production workspaces using a shared Sanity Studio codebase. The content model supported English and Spanish, reusable content components, audio-enabled experiences, and custom editorial tooling. A separate TypeScript publishing pipeline extracted content from Sanity, transformed it into application-ready documents, migrated selected media assets, validated output, and published it into Firebase.
My Contribution
Architected and maintained a multi-workspace Sanity Studio v3 platform with a structured content hierarchy spanning 30+ document types
Implemented bilingual English and Spanish content modeling using document-level internationalization
Built custom Studio tooling for inline audio preview, translation assistance, content search and replacement, and audio-production workflows
Developed safe TypeScript migration utilities for cloning, deleting, searching, replacing, backing up, and synchronizing connected content trees
Designed a Sanity-to-Firebase publishing pipeline with pure transformation layers, Zod validation, deterministic identifiers, batched Firestore upserts, dry-run support, and media migration
Configured automated Sanity Studio deployment and isolated environment workflows using GitHub Actions and separate datasets
Content Platform Architecture
One shared TypeScript Sanity Studio codebase across four managed workspaces: development, QA, staging, and production
Structured hierarchical documents for programs, playlists, sessions, lessons, activities, journals, and reusable input components
English and Spanish document-level localization with linked translation documents
Custom desk structure that keeps low-level input types out of top-level navigation
GROQ for content queries and relationship traversal
Translation assistance: admin-restricted helper for English and Spanish document workflows
Content operations: custom in-Studio find and replace, dry-run migration scripts, atomic deep clone, safe delete with reference checking and confirmation, plus dataset backup and synchronization
Sanity-to-Firebase Publishing Pipeline
Sanity CMS authoring and export
↓
Resolve content relationships and references
↓
Transform CMS documents into mobile application shapes
↓
Normalize identifiers, timestamps, media paths, and rich text
↓
Validate output with Zod
↓
Migrate selected audio assets to Firebase Storage
↓
Publish documents to Cloud Firestore with merge-based batched upserts
↓
Operator dry runs and repeatable re-execution
Publishing Pipeline Details
Primary supported approach: full snapshot synchronization from Sanity to Cloud Firestore and Firebase Storage
Partial incremental update scaffolding exists for selected content types; major structural changes require a full sync
Continuous hosted execution is not claimed by this portfolio — operators drive publishing with dry-run and re-run support
Sync Reliability
Deterministic identifiers and stable ID mapping across cloned content
Merge-based idempotent upserts and skip-existing media uploads
Pure transformation functions separated from I/O
Zod schema validation, dry-run modes, and offline snapshot processing
Vitest coverage for transformation behavior
Future improvements: deletion reconciliation, richer retry policies, CI for the syncer, and stronger observability
Engineering Challenges
Modeling a large interconnected content graph while keeping the editorial interface usable
Maintaining stable runtime identifiers when CMS content was cloned or reorganized
Transforming CMS-oriented schemas into application-specific Firestore documents
Supporting bilingual content and linked translation documents
Migrating and remapping media assets across storage systems while preventing schema and migration-script drift
Reliability, Privacy, and Security Considerations
Editorial and publishing systems for digital-health content require careful environment isolation, validation before publish, and conservative public documentation. This portfolio omits project IDs, dataset names, media identifiers, clinical content, and private screenshots.
Engineering Outcomes
Established a structured source of truth for application content
Reduced reliance on direct manual editing of runtime Firestore documents
Created repeatable publishing and migration workflows
Improved editorial safety through validation, dry runs, and atomic operations
Enabled isolated content work across multiple environments
Separated editorial authoring models from application runtime schemas
Supported consistent bilingual content management
Relationship to Other Projects
This project owns content authoring and publishing infrastructure
The Healthcare Mobile Application consumes the published Firebase content and focuses on React Native, Expo, authentication, state, notifications, and release delivery
Distinct from the Healthcare Data & ETL Platform, which handles batch operational data, Python transforms, and a MySQL analytical warehouse across serverless and container runtimes