Staff AI Full-Stack Architect

GlaxoSmithKline

Durham (NC)

On-site

USD 180,000 - 210,000

Full time

5 days ago

Be an early applicant

Get a reply from this employer

Your resume and cover letter, tailored to exactly what they’re hiring for.

Get past ATS filters

Job summary

GlaxoSmithKline in the United States is seeking a Staff Full Stack Engineer to set technical direction for the AI product stack across Enterprise AI, focusing on integration, context, and agent-driven development. You will define standards for API, platforms, APIs, and frontend engineering while guiding senior engineers toward architectural independence.

This hybrid role emphasizes LLMs, MCP protocols, observability, and scalable EAI platform decisions that improve reliability and developer

Qualifications

Bachelor's degree in Computer Science, Data Science, Statistics, Applied Mathematics or a related quantitative field, or equivalent practical experience.
10+ years of experience in applied machine learning (ML) or data science and full-stack engineering, including experience building and operating user-facing AI products.
Experience across the AI product stack, including modelling pipelines, evaluations, platforms and APIs, modern web frameworks, and data and cloud infrastructure.
Experience designing and operating production AI products, including large language model (LLM) selection and integration, orchestration and tool-use architectures, retrieval and context integration using protocols such as Model Context Protocol (MCP), safety, and cost, latency and quality optimisation.
Experience defining API, platform and frontend engineering standards for multiple teams, including interfaces designed for non-deterministic, agent-driven behaviour.
Experience building large language model operations (LLMOps), evaluation and monitoring infrastructure for production AI systems.
Experience defining and scaling agentic research and product workflows, such as auto-research, continuous evaluation and agent-driven application development.
Experience setting multi-quarter technical strategy across multiple engineering teams.
Master's or PhD strongly preferred, in Computer Science, Data Science, Statistics, Applied Mathematics or a related quantitative field, or equivalent practical experience.
Prior Staff, or equivalent, experience at an organisation shipping non-trivial AI products that combine modelling and full-stack delivery.
Defined organisation-wide engineering strategy, particularly for agentic research and product workflows, and seen it through to measurable outcomes.
Experience in a regulated industry, such as life sciences, where data integrity and auditability are mandatory.
Public technical contributions, such as talks, writing or open source, on AI-native, agent-augmented engineering
Experience defining API and platform architecture standards for multiple engineering teams and making platform decisions to improve reliability and developer productivity.
Experience with AI integration architecture, including large language model (LLM) selection, orchestration patterns and LLM operations (LLMOps)and implementing evaluation pipelines and AI observability practices across multiple teams.
Experience defining frontend engineering standards for AI products and designing non-deterministic, agent-driven user interfaces.
Experience defining LLMOps standards, including AI system observability, evaluation and governance and building evaluation and monitoring infrastructure for production AI systems.
Experience defining full-stack engineering standards, including standards for agent-generated code.
Experience developing developer tooling and shared component strategies.

Responsibilities

Set multi-quarter technical direction for the AI product stack across EAI: modelling workflows, platforms and APIs, AI-powered interfaces and the agentic systems that connect them.
Define EAI's agentic research practice for continuous experimentation and auto-research, evolving production models, and using agents to drive research and application iteration.
Set the operating model for agent-generated engineering work: what digital workers build and change on their own, what requires human approval, which evaluation gates their work must pass and how incidents evolve.
Set the AI-assisted development standard: the tooling, workflows and practices for coding and research agents, with measurable impact on velocity and quality.
Define API and platform architecture standards adopted across EAI engineering teams, covering contracts, authentication, versioning and non-functional requirements.
Drive architectural coherence across modelling pipelines, evaluations, retrieval, agent orchestration, application services, frontends, data and observability, without becoming a bottleneck.
Shape EAI's AI integration architecture: LLM selection, orchestration patterns, retrieval and tool use, and LLMOps standards.
Own how EAI products and agents draw on the governed context layer over standard protocols such as MCP, with consistent patterns for retrieval, permissions and provenance.
Define EAI's frontend engineering standards and shape how AI capabilities reach users across the portfolio.
Lead EAI's approach to AI-native UX: designing for non-deterministic, agent-driven interfaces, with confidence signals, error states, provenance and human-override paths that earn user trust.
Define EAI's LLMOps standards for observability, evaluation and governance of production AI systems, and lead enterprise-scale evaluation and monitoring infrastructure.
Own long-term system health across reliability, performance, cost, security, model behaviour, safety and abuse, as EAI's authority on production AI system health.
Define full-stack engineering standards adopted across EAI pods, written to be followed by humans and digital workers alike.
Shape EAI's developer tooling and shared component strategy, leading foundational investments such as evaluation frameworks, agent runtimes, experiment management, internal SDKs and shared components that reduce duplication and change how teams build.
Develop senior engineers towards greater architectural independence, and partner with research, product, design, data engineering and platform teams on the hardest product problems.
In the first 12 months, the person in this role will have:
Set and begun executing a multi-quarter technical strategy for the AI product stack across EAI.
Established the operating model and standards under which engineers and digital workers build across platforms, APIs and interfaces, and led at least one platform investment that changed how teams build.
Defined the integration and orchestration architecture through which EAI products and agents draw on the governed context layer, with evaluation and observability standards in production.
Defined EAI's approach to agentic research and AI-assisted development, with measurable adoption and impact on velocity and quality.
Raised the architectural independence of EAI's senior engineers through standards and coaching.