Staff, Data Engineer

GSK•Philadelphia, PA

1d•Hybrid

About The Position

The Staff Data Engineer is accountable for implementing and evolving the governed context layer used by people and digital workers across GSK's Enterprise AI (EAI), and serves as EAI's internal authority on data architecture for AI. The context layer makes data, meaning, permissions and history retrievable when needed, connecting enterprise knowledge, structured data, business vocabulary and modeled ontologies across data platforms and AI experiences. The role defines the architecture, standards, frameworks, contracts and guardrails within which digital workers generate, test, deploy, monitor and remediate data pipelines. It shapes AI data infrastructure and lifecycle practices, responsible data governance, and the data platform architecture, while ensuring shared components are reliable, performant, cost-effective and secure. The role also drives data engineering standards across EAI and develops senior engineers' architectural capabilities. This role is based in the United States and follows a hybrid working model, with on-site collaboration roughly two days per week. Remote work outside this arrangement is not available.

Requirements

Bachelor's degree in Computer Science, Data Engineering or a related field, or equivalent practical experience.
10+ years of data engineering experience, including defining data architecture for production AI or machine learning systems.
Experience building AI data infrastructure, including feature stores, vector databases, retrieval architectures, streaming data pipelines and evaluation datasets.
Experience building context, metadata or semantic infrastructure used by AI systems, such as data catalogues, business glossaries, knowledge graphs or ontologies, exposed to AI agents through protocols such as Model Context Protocol (MCP).
Experience designing and operating AI agents or large language model-powered systems within data workflows, including evaluation, guardrails and human-in-the-loop controls.
Experience defining data quality, lineage and governance standards for AI pipelines across multiple data domains, including privacy, consent, PII handling and auditability.
Experience with cloud data platforms and infrastructure as code, including setting engineering standards.

Nice To Haves

Prior Staff, Principal or equivalent experience in an organization operating a non-trivial data platform for AI.
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 data architecture for AI, context engineering or agent-operated data engineering.
Experience defining pipeline architecture standards for enterprise AI and multiple AI initiatives.
Experience implementing and evolving governed context and semantic layers used by people and AI systems at run time.
Knowledge of data architecture for AI, including feature engineering, vector infrastructure, streaming and evaluation data design.
Experience with versioning, lineage, evaluation dataset governance and model data lifecycle management.
Experience designing cloud data platform architecture, applying infrastructure-as-code and platform-as-product practices, and building reusable components for multiple AI initiatives, engineering teams and AI systems.
Experience mentoring senior engineers, supporting architectural decision-making and advising on complex data challenges.

Responsibilities

Implement and evolve the governed context layer as a product spanning enterprise knowledge, structured data, business vocabulary and modeled ontologies, usable by humans and digital workers at run time.
Build context-layer capabilities, including cataloguing structured and unstructured data, AI-assisted metadata enrichment, use-case-specific context design, secure agent access through standard protocols such as MCP, and routing across sources.
Make certified business meaning machine-readable, with governed definitions, ownership and lineage for data returned by digital workers.
Define architecture standards and shared frameworks for digital workers to generate, test, deploy and monitor pipelines, reducing integration costs for new AI initiatives.
Set the operating model for agent-generated engineering work, including autonomy boundaries, human approvals, quality evaluation and incident escalation; lead incidents digital workers cannot resolve through permanent resolution.
Shape EAI's AI data infrastructure strategy across feature engineering, vector and retrieval infrastructure, streaming data flows and evaluation data design; build shared production components for multiple AI initiatives.
Define EAI's MLOps data practices, including versioning, lineage, drift monitoring, evaluation dataset governance and model data lifecycle management, including context and evaluation data used by digital workers.
Define responsible data governance standards for AI pipelines covering privacy, consent, PII, auditability and bias controls, and encode controls in the governed context layer.
Shape data quality and lineage across multiple domains, with measurable outcomes in error rates, downstream AI performance and audit readiness.
Design EAI's data platform architecture using infrastructure as code and platform-as-product practices so data and AI projects can start consistently and quickly.
Build shared production platform components and own their reliability, performance, cost and security.
Define and drive adoption of data engineering standards for code, documentation and review that can be followed by people and digital workers.
Develop senior engineers toward greater architectural independence and partner with forward deployed engineers, ML engineers, platform engineers and domain experts on production AI data requirements.