At Florida Child Care Connect—DXP platform for the Florida Department of Children and Families—a US state-government agency—I joined a cross-functional team handling development, design, DevOps, PR reviews and task management to deliver a highly secure, data-driven platform. I architected a resolver-level RBAC system, slashing unauthorized access by 90% and boosting API throughput 15%, and built a custom TypeScript-generic DynamoDB ORM layer with Joi validation that accelerated CRUD development by 60% while ensuring 99.9% data integrity. I engineered a serverless analytics pipeline in AWS Lambda for real-time aggregation of 10K+ daily events—delivering sub-second dashboard queries and 95% faster decision velocity—and implemented county-level geospatial engagement metrics with 99.9% algorithmic accuracy to inform targeted content strategies. To support global use, I added English/Spanish GraphQL locale support and optimized filter expressions to cut payloads 30%. On the frontend, I modularized Redux Toolkit slices and Apollo Client caching to reduce network payloads 45%, accelerate UI responsiveness 60%, and shrink bundle sizes 35% via custom Webpack tree-shaking and gzip compression. All work, from Figma prototypes to live demos and social previews, is documented in Notion.

Media

Project Points Earned

  1. Architected Enterprise-Grade RBAC with JWT-Based Context Authentication – Designed a fine-grained, resolver-level role-based access control system in GraphQL using early-exit guards and JWT secrets, slashing unauthorized access incidents by 90% and boosting API throughput by 15%.
  2. Built Custom DynamoDB ORM Framework with TypeScript Generics & Joi Validation – Engineered a class-based ORM layer for AWS SDK v3, supporting 10+ entity models with auto marshalling/unmarshalling and Joi schema enforcement, achieving 99.9% data integrity and 60% faster CRUD development.
  3. Engineered Real-Time Analytics Engine with Lambda & In-Memory Aggregation – Developed a serverless pipeline in AWS Lambda to process 10k+ daily events with O(n) complexity, implementing memory-efficient aggregation and date-range filters to deliver sub-second dashboard queries and improve executive decision velocity by 95%.
  4. Architected Geospatial Intelligence & User Behavior Analytics – Leveraged AWS Lambda and TypeScript to compute county-level engagement metrics with complex statistical algorithms (99.9% accuracy), driving a 40% lift in targeted content strategies for regional market expansion.
  5. Developed Flexible Time-Series Analytics with Dynamic Granularity – Created a multi-resolution engine (8-hour, 24-hour, 7-day, 30-day, yearly) using adaptive sorting and locale-aware parsing, reducing historical data processing time by 60% across 12-month datasets.
  6. Delivered Predictive BI Dashboard with Chart.js & Advanced Metrics – Built interactive KPI dashboards featuring top-10 resource rankings, category performance, and floating-point precision analytics, driving a 40% boost in content engagement through actionable insights.
  7. Implemented Resource Recommendation & View-Tracking Engine – Designed a county-aware recommendation system combining view-count telemetry and weighted priorities, increasing click-through rates by 35% and surfacing top content dynamically per region.
  8. Engineered Multilingual API Infrastructure for Global Expansion (EN/ES) – Added GraphQL locale arguments and DynamoDB attribute projections for English/Spanish, expanding addressable users by 40% while reducing irrelevant payloads by 30%.
  9. Optimized Frontend State Management via Redux Toolkit Domain Architecture – Modularized state into seven feature slices with Redux Persist and middleware orchestration, cutting state complexity by 40% and accelerating UI responsiveness by 60%.
  10. Built GraphQL-First API Integration with Apollo Client Caching & Batching – Implemented InMemoryCache policies, query batching and type-polymorphic fragments, reducing network payload by 45%, API calls by 35%, and fetch latency by 50%.
  11. Streamlined Webpack CI/CD Pipeline with Tree-Shaking & Gzip Compression – Customized production builds with lodash tree-shaking, bundle analysis and gzip minification, shrinking bundles 35% and improving initial load Cumulative Layout Shift scores by 25%.
  12. Optimized DynamoDB Read Patterns with Strategic GSI & Cursor Pagination – Refactored queries to leverage Global Secondary Indexes, projection expressions and cursor-based pagination, cutting read latency by 40%, data-omission errors by 50%, and RCU consumption by 20–40%.