Frontend Developer Mid

We are seeking a Frontend Developer to build and own the web frontend of a TMS dashboard with integrated AI actions. This is a complex, data-heavy application requiring strong React and TypeScript skills. The role involves rebuilding a frontend against an existing backend mapping to API contracts not designed by the candidate and building AI-integrated UI features including streaming chat, OCR review interfaces, and AI action audit trails. Key Responsibilities • Build and maintain complex, data-heavy dashboard UIs using React with hooks, context, and component architecture • Implement role-based UI what users can see and do is determined by their user type • Design and develop complex multi-field forms with validation, conditional fields, and server-side error handling • Integrate with existing backend API contracts (REST) mapping to endpoints not designed by you • Implement state management at scale using Redux, Zustand, React Query, or a combination • Build real-time UI features using WebSockets or Server-Sent Events for live shipment status updates without page refresh • Develop a streaming chat interface handling LLM token-by-token rendering via SSE or WebSocket • Build AI action receipt UI auditable display of what the AI did, what data it changed, and a path to review or reverse it • Create OCR review interfaces: scanned document alongside AI-extracted fields with confidence indicators and editable correction fields before committing to the database • Implement optimistic UI patterns for AI actions with clean rollback on failure • Build file upload and document management UIs: drag and drop, upload progress, preview, and status tracking through the OCR pipeline • Build data visualizations: charts, maps, and status boards for operational monitoring • Ensure responsive design across all pages many pages will need to be created beyond existing designs • Handle async UI patterns throughout: loading states, error states, and optimistic updates

Qualifications: Must-Have (Non-Negotiable) • 3+ years React development with a portfolio of complex, data-heavy dashboard applications — React with hooks, context, and component architecture (not just JavaScript, not just "frontend") • Strong TypeScript — the TMS dashboard with AI actions is too complex for untyped JavaScript • Experience rebuilding a frontend against an existing backend, mapping to API contracts you did not design • Deep experience with state management at scale: Redux, Zustand, React Query, or a combination • Strong proficiency with a component library (Material UI, Chakra UI, Ant Design, or equivalent) • Experience building role-based UI • Strong understanding of form architecture in React: complex multi-field forms with validation, conditional fields, and server-side error handling • Responsive design that actually works — simple yet effective. Able to create pages beyond existing UI designs. • Strong understanding of asynchronous UI patterns: loading states, error states, and optimistic updates Strong Preference • Experience building dashboard-heavy applications: data tables, filters, status boards, and real-time updates • Experience building multiple frontends that share one backend • Experience with WebSockets or SSE for real-time data — shipment status updates push to the UI without refresh • Experience with form-heavy UIs connected to real business data: load entry, driver assignment, and document uploads • Experience with data visualization: charts, maps, and status boards for operational monitoring • Has built file upload and document management UIs: drag and drop, upload progress, preview, and status tracking for documents going through an OCR pipeline AI-Specific Requirements • Has built a web-based streaming chat interface — handling SSE or WebSocket streaming from an LLM API, rendering progressively as tokens arrive • Experience designing and building AI action receipt UI: clear, auditable display of what AI did in the system, what data it changed, and a path to review or reverse it • Has built OCR review interfaces: AI-extracted fields from a scanned document alongside the original document image, with field-level confidence indicators and editable fields for human correction before committing to the database • Understands optimistic UI patterns for AI actions — show the expected result immediately while the backend confirms, and handle rollback cleanly if it fails • Experience with AI process status UIs: progress indicators, background task monitoring, and notification patterns for long-running AI processes like batch document processing