Senior Full Stack Engineer, AI Platform

Location: Work from home (Pennsylvania) Shift: Days (United States of America) Scheduled Weekly Hours: 40 Worker Type: Regular Exemption Status: Yes Job Summary: Geisinger is operationalizing AI at scale — moving past pilots into a portfolio of production AI capabilities serving 70+ programs across clinical care, operations, the health plan, and pharmacy. None of that scales without strong internal applications: the dashboards, evaluation tools, and developer portals that program teams, governance stakeholders, and platform engineers actually live inside every day. This role is where those applications get built. The Senior Full Stack Engineer owns the end-to-end product quality of every user-facing application the AI Platform ships — design system to component library, backend-for-frontend APIs to telemetry, role-based access to load-time budgets. Whenever a program lead opens a platform dashboard, a governance reviewer opens an evaluation tool, or a developer hits an internal portal, this is the engineer whose work decides whether they trust what they see and come back tomorrow. The role exists because internal AI tooling has stopped being a nice-to-have. The pace of program delivery is now bounded by how quickly teams can read platform data, configure platform capabilities, and self-serve onboarding into the platform's standard paths. We need a senior engineer who treats internal applications as a real product surface and who is opinionated about what good looks like. Job Duties: Why This Role Matters:   The AI Platform is an enabling team that builds reusable capabilities for every AI program at Geisinger. Most of those capabilities have a user-facing surface — dashboards, configuration interfaces, review tools, developer portals and each of those surfaces is the moment a program team, a governance reviewer, or a platform engineer decides whether the platform is helping them or getting in their way. Slow, ugly, or inaccessible internal applications quietly kill adoption and adoption is what the platform is measured on. The application portfolio will evolve as platform capabilities evolve; the constant is that someone has to own the bar for what those applications look and feel like.   This is an individual contributor role with a broad surface area. You are the senior frontend voice for the AI Platform, the steward of the design system, and the engineer accountable for whether the applications actually feel good to use. What You Will Own:   The frontend architecture and the shared design system / component library used across every AI Platform application — visual and interaction consistency.   RBAC-aware interfaces — making sure program teams, governance stakeholders, leadership, and platform engineers each see the right data and controls based on their role   The backend-for-frontend layer — API patterns, data contracts with MLOps, and the seams where the UI meets the platform   Application-level testing strategy — unit, integration, and end-to-end suites that the platform's deployment verification pipeline can trigger automatically   Usage telemetry and analytics — instrumenting every application for adoption, feature engagement, error rates, and load times, and surfacing the insights that inform the platform roadmap   Frontend performance — bundle optimization, caching, lazy loading, and load-time budgets for internal applications where slow UIs erode trust   UX quality gates — defining and enforcing the bar that no application ships below Shape of the Work   This role lives at three altitudes:   With the design system (hands-on build).   Own the shared component library and the patterns underneath it. Build accessibility in from the start, not as a remediation pass. Make the design system something other engineers want to consume because it makes their work faster, not because they're forced to.   With application teams (collaborative delivery).   Build the user-facing applications in the platform's portfolio alongside the MLOps engineers who own the backend logic. Negotiate API contracts, RBAC implementations, and data shapes with the integration engineer and the platform engineer so the UI and the backend evolve together. Today's portfolio is one set of applications; tomorrow's will be another, and you'll own the interface layer through the transitions.   With the data the applications generate (product instinct).   Wire telemetry into every surface and read it. Page views, feature engagement, drop-off, error rates, load times. Bring evidence to the platform's roadmap conversations: which features get used, where users get stuck, which application surfaces deserve more investment and which ones to retire. Use what you instrument to drive what you build next.   Today's Application Portfolio:   The platform team currently maintains a small, growing set of internal applications. Examples of what's on the floor right now:   An   AI Scorecard   dashboard displaying per-initiative status across Performance, Adoption, Outcome, and Equity pillars   An   LLM-as-Judge evaluation tool   for configuring evaluation criteria, reviewing judge outputs, and tracking eval results over time   Developer portals and internal tooling   — documentation sites, onboarding flows, and self-service interfaces for the platform's standard delivery path   Expect the portfolio to grow and shift as new platform capabilities come online. The role is the application layer of the AI Platform, not these specific applications.   Key Technologies:   React 18 (frontend framework)   Mantine and/or MUI (component library foundations for the design system)   FastAPI (backend-for-frontend APIs)   SQLAlchemy (ORM / database access)   Docker (containerization)   Jest, React Testing Library, Playwright or Cypress (testing)   OpenTelemetry and application-level metrics (telemetry and observability)   Collaboration Poin