Senior Staff AI Application Engineer

Job description

Job Description SummaryWe are looking for a senior full-stack engineer who has shipped production applications, and who has owned software that other engineering teams depended on — a framework, SDK, shared library or internal developer platform — living with an abstraction decision long enough to have had to correct it while consumers depended on it. Establishing engineering standards through automation rather than documentation should be instinctive. You will be the senior technical owner of GE HealthCare's internal AI engineering platform: the shared libraries, standardised interfaces, code generation tools and automated verification our data scientists and engineers use to take AI solutions into production. It exists to deliver four outcomes for the company: • reduce time from working prototype to production from months to days; • allow data scientists to ship production-grade AI without expertise in cloud infrastructure, deployment, application security or front-end development; • make security, compliance and consistency automatically enforced properties of every solution, rather than outcomes dependent on scarce expert review; • allow underlying technologies to be replaced — model provider, agent framework, user interface approach — without rewriting the solutions built on them. Your accountability is the platform's architecture and public interfaces, the automated verification that enforces our engineering standards, and the versioning and migration approach that keeps it dependable as it evolves beneath a growing number of applications. The role also includes building AI applications end to end yourself — front end, back end and agents — where that is the fastest way to prove a platform capability or unblock a priority use case; the emphasis at this level sits firmly on the platform, with application work serving it rather than competing with it. This is a hands-on senior engineering role with substantial architectural responsibility and organisational influence — not a management position, and not a research position. Job DescriptionGE HealthCare's Chief Data and Analytics Office is delivering a portfolio of strategic AI programmes across Commercial, Finance, Supply Chain, Quality, Manufacturing and Operational Excellence. The problem this team addresses is structural. Most people creating AI solutions here are data scientists, and it would be neither reasonable nor efficient to require deep infrastructure, security and front-end expertise of them. Every solution must nonetheless reach production as a secure, reliable, observable, supportable enterprise system, and remain so through years of change. That is the central design constraint: the platform and its automated safeguards must substitute for a depth of software engineering review that is not available at the scale and pace we require. The failures a strong platform engineer prevents — undetected interface drift, poorly chosen abstractions that become permanent, over-broad access permissions, configuration that cannot be reviewed or reversed — are silent, and present months later as an outage, an audit finding or a security incident rather than as a broken feature. Core ResponsibilitiesArchitecture and design Define the platform's architecture and public interfaces: what is offered, what is guaranteed, what is private, and what a version number promises to dependent applications. Design isolation boundaries so each significant third-party technology — agent framework, model provider, message queue, data source, UI rendering approach — is confined to a single replaceable component, with automated checks making that a verified property rather than a documented intention. Validate each in practice, for example by running the same agent unmodified on two frameworks (LangGraph and AWS Bedrock AgentCore, say) with equivalent evaluated behaviour. Design the interfaces between roles and components as machine-readable specifications with automated verification, so no two teams need to read each other's code to work together reliably. Hold the boundary between declarative configuration and executable code — configuration states what, code implements how — preventing the familiar progression by which a configuration format acquires conditional logic and becomes an undocumented programming language with no type checking, debugger or tooling. Record architectural decisions formally (options rejected, trade-offs accepted, revisit conditions), and decline unnecessary abstraction in writing and with reasoning. A significant part of this role's value lies in shared capability deliberately not built. Engineering delivery and long-term ownership Lead the highest-leverage components hands-on: the agent execution layer, the secure data resolution layer, and the configuration-driven user interface system. This is a building role, not a supervisory one. Build selected AI applications end to end — front end, back end and agents — where that is the fastest way to prove a platform capability or unblock a priority use case, and judge what should then be harvested into the platform and what should remain specific to that use case. Establish automated verification before the practice it enforces is adopted — a standard that is not automatically enforced does not, in practice, exist — including automated quality evaluation of AI behaviour in the release path, so degraded output prevents release and a change of model or framework can be shown to have preserved behaviour. Own backward compatibility and the migration path. The platform will change continuously while dozens of solutions depend on it; making that change inexpensive and safe for consumers is a primary responsibility. Prevent version divergence across applications and shared components; run deprecation properly — announce, migrate, remove — and maintain visibility of the platform's own engineering health, acting before it is raised. Interfaces with adjacent disciplines Define and maintain the interfaces between this team and adjacent specialist functions — infrastructure, security, data — as version-controlled specifications with automated validation rather than request queues or standing meetings. These cover what an environment provides, what an infrastructure component accepts, what a service requires to run and is permitted to do, what must pass before release, what artefact is deployed and how it is promoted, and what telemetry and service level commitments apply. Anticipate and act on the recognised ways such boundaries deteriorate: a request queue re-forming, resources created outside the sanctioned path, release throughput constrained by a single function, divergent standards emerging, or incidents beginning with a dispute about ownership. Represent the platform's architecture and trade-offs to enterprise architecture, information security, risk and senior leadership, in terms appropriate to each audience. Technical leadership Mentor engineers and data scientists moving into platform and application engineering, and establish a review culture in which human review addresses design judgement and business fit, because formatting, type correctness, security fundamentals and interface compliance are handled automatically. Author the guidance, reference implementations and worked examples that let colleagues and AI coding assistants work productively without direct supervision, and set the standard for responsible use of those assistants under named human accountability. Experience RequirementsBachelor's degree in Computer Science, Software Engineering or a related field, or equivalent demonstrable practical experience. An advanced degree is welcome but not required. Minimum 8 years of professional software engineering experience, including demonstrable ownership of a shared library, framework, SDK or internal developer platform relied upon by other engineering teams. This is the essential requirement: not a record of delivering many applications,