Senior Fullstack AI Engineer

The role: An AI-native application that turns requirements into working technical products: it gathers structured specifications such as data contracts, schemas, standards and control definitions alongside messy business requirements, assembles the relevant context, spots what is missing or contradictory, generates code and artifacts, and then validates its own output against the original requirements and the organisation's controls. Your focus is the engineering around the model, which is where these systems are actually won or lost.

The generation and validation loops. The harnesses through which models reach context, tools and execution environments. And the human-AI interaction itself, because how the application asks, clarifies and hands back control determines whether it understands intent or merely appears to. This is a hands-on individual contributor role. You would work with other senior engineers whose expertise complements yours, sharing the technical and architectural decisions rather than owning them alone.

The environment: A multidisciplinary team building something genuinely new, inside an organisation with real enterprise requirements. Security, privacy, identity and access management, accessibility, auditability and operational reliability are not obstacles to route around here; they are part of what makes the output trustworthy enough to use. You would collaborate with architecture, data engineering and platform specialists, which means a good deal of the work is finding the responsible way to move fast rather than choosing between the two.

What you'll do: Design human-AI interactions that gather requirements, retrieve context, resolve ambiguity, and let users review and refine what was generated. Build an enterprise-compliant web application: interface, APIs, application services, integrations and data model. Engineer the harnesses that connect models with context, tools, execution environments, quality controls and user feedback. Develop iterative loops in which AI generates, tests, validates, corrects and explains its own output.

Integrate quality checks, automated tests, CI/CD feedback and organisational controls into the generation process. Maintain traceability between requirements, retrieved context, generated artifacts, validation outcomes and user decisions. Contribute to the security, observability, maintainability and operational readiness of the application.

Technical ground: Core: full-stack web engineering (interfaces, APIs, services, integrations, data models) · AI harness design · context engineering · iterative generation and validation loops · human-AI interaction design.

Also in the mix: AI-assisted code generation and secure execution · evaluation harnesses and guardrails · requirements traceability and policy-as-code · data contracts, schemas and metadata · containers and CI/CD.

What you bring: Must-haves: Senior-level experience designing, building and operating enterprise web applications. Strong full-stack engineering across user interfaces, APIs, application services, integrations and application data models. Experience designing human-AI interactions that gather context, resolve uncertainty and support review of generated output. Experience engineering AI harnesses and the loops around them: generation, testing, evaluation, correction and escalation to a human.

A real grasp of context engineering, including how context is gathered, retrieved from enterprise sources, assembled, maintained and evaluated. Experience bringing software engineering practice into AI workflows: automated testing, CI/CD, data-quality checks, observability and controls.

Particularly valuable: AI-assisted code generation, and the secure execution and validation of generated code. Building evaluation harnesses, guardrails or quality feedback mechanisms. Requirements traceability, policy-as-code, automated control checking or approval workflows. Familiarity with data contracts, schemas, metadata, data-quality controls or data-product concepts. Cloud-native delivery, containers and CI/CD. You understand that the model is one component of an AI-native application, and that reliable outcomes come from how context is gathered, how the application talks to people, which tools and controls exist, and how feedback feeds the next iteration.

You know what enterprise requirements mean in practice and why they exist, you will question an assumption constructively, and you can turn an incomplete requirement into something working and testable. You take ownership of hard problems without casting yourself as the only technical authority in the room.