Kevin Diesenberg

• Maintained a server-rendered React and TypeScript platform on Arc Fusion and Node.js across 6 broadcast station sites serving 1.05M daily users.
• Built an internal Trending Stories Dashboard that combines algorithmically trending stories with editor-submitted content, giving editors control over creating, updating, and expiring placements distributed through an API consumed by branded apps for display in the main navigation menu and in the main header carousel I built.
• Integrated the Vercel AI SDK into editorial workflows to automate story summary generation, reducing manual content-operations time and supporting faster publishing.
• Adopted agentic AI coding tools (Cursor, Claude Code) across feature development and code review, accelerating delivery while maintaining type safety and review standards on a production codebase.
• Developed standalone internal tools with Vite, TanStack (Query, Table, Form), and shadcn/ui, backed by SST, AWS Lambda, and DynamoDB to support rapid iteration and data normalization across multi-station editorial workflows.
• Reduced redundant client-side API calls by 20% by improving caching, request deduplication, and conditional fetching with SWR and Axios.
• Built secure authentication and subscription-gated experiences using Amazon Cognito, the Arc Identity SDK, and Duo SSO in a server-rendered environment.
• Expanded shared UI patterns by composing themed styled-components over Headless UI and Radix UI primitives, improving accessibility through keyboard support, reduced-motion behavior, and reusable dialog patterns.
• Added Sentry for production monitoring and contributed to a navigation state refactor across 6 sites that reduced prop drilling and improved shared query patterns.

• Rigged lighting, sound, and staging for 1,000+ live events across arenas, theaters, and convention centers, including 3 presidential inaugurations

• Operated backcountry water systems serving 50,000+ park visitors

Steamy Noods

Crowdsourced Vegan Noodle Soup Discovery Platform

• Sole developer of a full-stack TanStack Start application with hybrid SSR/CSR, type-safe server functions, route-based data loading, auth middleware, and admin role checks.
• Designed a reputation-weighted vegan verification system with reusable Drizzle SQL fragments for aggregated queries.
• Built advanced restaurant search with multi-filter queries, computed fields, full-text search, cursor-based pagination, and indexed PostgreSQL tables.
• Added optimistic updates with rollback handling, session-based auth, IP rate limiting, and transactional writes to protect data integrity.

OR Utilization Simulator

Operating-Room Scheduling Tool for an Orthopaedic Surgery Practice

• Built a local-first single-page app, currently used by a surgical practice, that models OR utilization across multiple staffing scenarios (1 OR serial; 2 ORs + 1 CRNA; 2 ORs + 2 CRNAs) with automatic constraint enforcement.
• Designed a deterministic capacity-packing engine that schedules cases within each block's time window while respecting shared-resource constraints per scenario, with user-pinned case ordering via drag-and-drop.
• Implemented month/week/day Gantt calendar views with drag-and-drop OR blocks, a CRUD case library with per-phase time estimates (turnover, anesthesia/draping, surgeon-critical, closing), and composable case-mix presets driving the packing algorithm.
• Built a real-time insights rail (CRNA hours vs. guaranteed minimum, OR efficiency %, per-surgeon case hours) over Zustand state persisted to localStorage with versioned schema migrations (v1 to v3).
• Covered the packing engine, store migrations, and scenario constraint logic with 74 unit and integration tests in Vitest.