Description
Enterprise AI (EAI) builds AI products that combine real modelling work with user-facing applications and ships them to production. Full Stack Engineers move across data, models, platforms, APIs and interfaces in a single workflow.
The Full Stack Engineer II independently builds and ships AI-integrated features, working closely with GSK business teams to prototype against real data and workflows at a days-to-weeks cadence.
Key Responsibilities:
- Delivery & Experimentation:
- Own use cases from problem framing through working prototype and first production version, at a days-to-weeks cadence and without close supervision.
- Qualify incoming asks: challenge underspecified problems, assess feasibility and define scope before committing to work.
- Work closely with business teams, prototyping against real data and workflows and building trusted relationships with product owners and engineering leads.
- Agent-Operated Engineering:
- Build with digital workers as the default mode of engineering: direct them, validate what they generate and course-correct at engineering speed.
- Design and orchestrate multi-agent workflows end-to-end, and judge when agentic approaches add value over simpler alternatives.
- Platforms & APIs:
- Design and ship backend services independently with observability, clear contracts and compliance with non-functional requirements.
- Choose API design patterns that suit the consumer and iterate on them based on real integration feedback.
- AI Integration & Context:
- Build AI-integrated features independently, including RAG pipelines, tool use and multi-step LLM workflows that draw on the governed context layer with the right permissions and provenance.
- Test AI-powered features with structured evaluation frameworks before and after release.
- AI-Native Experience:
- Deliver complete UI features with measurable user engagement, bringing user feedback directly into build decisions.
- Build AI-powered interfaces with appropriate transparency: confidence signals, error states, provenance and human-override paths.
- LLMOps & System Health:
- Build evaluation pipelines and LLM-specific monitoring for your features, including output quality signals, model drift and guardrail compliance.
- Respond to AI production incidents with structured diagnosis, and see them through to a fix.
- Standards & Reusable Components:
- Build reusable components, SDKs, agents, prompt templates and reference architectures adopted by engineers outside your own pod, and drive consolidation where you find duplication.
- Raise team standards through code review, documentation and direct feedback, and resolve systemic quality issues in the codebase.
Required Experience:
- 10+ years of full-stack or AI/ML engineering experience, including production deployments.
- Demonstrated ability to take ideas from sketch to working software at speed, independently and under time pressure.
- Strong proficiency in Python and at least one of TypeScript, JavaScript, Go or similar, with experience designing and shipping backend services and complete UI features.
- Hands-on experience with GenAI patterns: agents, multi-agent orchestration, RAG, tool use, prompt engineering and evaluation, including evaluation pipelines and LLM-specific monitoring in production.
- Daily use of AI coding agents as the primary mode of development, with sound judgement on what to accept, change or reject.
- A track record of raising team standards through code review and documentation, and of contributing reusable engineering patterns that others adopted.