Role Overview

A senior engineering leadership role responsible for the technical direction, hands-on delivery, and production scaling of AI solutions across enterprise enabling functions — including HR, Finance, Procurement, Legal, Audit, Compliance, and Business Development.

This is a builder-leader role. The Engineering Director combines deep hands-on AI engineering — designing and shipping multi-agent systems, RAG pipelines, and governed AI applications — with the technical leadership required to drive a team from opportunity identification through to production deployment. They bring a rare and deliberate combination: the ability to move from idea to working proof-of-concept in days, alongside significant depth in AI governance, operational resilience, and regulatory compliance — not as adjacent knowledge, but as a core professional discipline that shapes how they build, assess, and operate AI systems.

The role sits within the Enterprise AI function and works in close partnership with enterprise technology, data engineering, technology governance, legal, information security, and functional stakeholders to deliver AI that is production-grade, auditable, and compliant from the first commit — not retrofitted at the end. Given geographic considerations, the role carries particular responsibility for navigating multi-jurisdictional data sovereignty, regulatory divergence, and cross-border AI governance — ensuring systems are defensible under all applicable regulatory regimes.

Context

Enabling functions — HR, Finance, Procurement, Legal, Audit, and Compliance — govern how an organisation hires, contracts, spends, reports, partners, audits, and maintains compliance. They represent high-value AI opportunities and high-consequence environments — where outputs carry regulatory, financial, and reputational weight. Realising value at scale requires engineering leadership that can navigate complex data landscapes, build for reuse, and embed governance, human oversight, and operational resilience into architecture decisions from the outset.

These AI applications do not exist in a vacuum. Each system must be governed — classified, registered, monitored, auditable, and defensible to regulators, auditors, and internal oversight functions. The governance and resilience challenge is twofold: building AI systems that are themselves resilient and well-governed, and ensuring the frameworks, processes, and controls that surround those systems are robust, proportionate, and continuously maintained. The role demands someone who has operated at this intersection for a significant portion of their career — not someone encountering governance as a new discipline.

This role is designed for an engineer who has already built AI applications inside a large, regulated enterprise, who has demonstrated experience delivering AI solutions across multiple enabling functions (e.g., HR, Finance, Procurement, Legal, Audit), who has significant experience governing AI systems and embedding operational resilience disciplines around them, and who treats regulatory requirements as architecture decisions — not compliance checkboxes.

Key Responsibilities

1. Technical Direction & Architecture

• Lead the engineering roadmap for AI across enabling functions, aligning architecture, delivery sequencing, and capability development to business priorities across HR, Finance, Procurement, Legal, Audit, and Compliance
• Set architectural direction for scalable, governed AI platforms — designing for modularity, cross-functional reuse, and compliance from the outset
• Make high-consequence technical decisions on architecture, build-vs-buy, model strategy (foundation models, fine-tuning, multi-provider orchestration, RAG), and integration patterns
• Drive platform thinking over project thinking — building shared components, reusable agent patterns, and common governance instrumentation that accelerate delivery across the portfolio
• Ensure architecture accounts for data sovereignty requirements — model routing, data residency, and hosting decisions that respect jurisdictional boundaries and cross-border transfer requirements
• Shape investment cases for senior stakeholders, articulating engineering decisions in terms of scalability, risk, regulatory defensibility, and value creation

2. Hands-On AI Engineering & Delivery

• Design and ship multi-agent LLM architectures across multiple model providers, choosing models against product requirements and compliance constraints — including sovereignty-aware routing through region-specific infrastructure where required
• Build RAG pipelines over real enterprise corpora with named single-purpose agents, hallucination guards before any user-facing output, and immutable audit logging at every stage
• Use AI-assisted development tooling to compress delivery from months to days, while keeping architecture decisions, model routing, and guardrails under deliberate human control
• Lead technical design for complex solutions spanning enabling functions — HR policy automation, contract risk scoring, procurement analytics, compliance monitoring, financial forecasting, audit analytics, and document intelligence — with governance built in from the first build
• Ensure rapid experimentation capability with clear engineering gates between proof-of-concept, pilot, and production — measuring against real data and real success criteria, not mock demos

3. AI Governance & Regulatory Compliance

This is a defining pillar of the role. The organisation requires an engineering leader with significant, demonstrated experience in AI governance — someone who has designed governance frameworks, built governance tooling, and operated in governance roles — not simply complied with governance requirements set by others.

• Governance architecture: Design and operate the governance structures that surround AI applications — classification and tiering, risk assessment, model registration, approval workflows, ongoing monitoring obligations, and decommissioning criteria
• Regulatory compliance (multi-jurisdictional): Ensure systems meet requirements under applicable data protection laws, AI-specific regulations (e.g., EU AI Act risk classification, emerging national AI frameworks), and sector-specific operational resilience expectations — navigating divergence and maintaining defensibility under multiple regimes
• Data sovereignty: Design data-residency and model-routing approaches that respect adequacy arrangements, data-transfer mechanisms, and sovereignty constraints — ensuring processing is appropriately separated by jurisdiction where required, with sovereign model options (e.g., region-specific cloud deployments, local model hosting)
• Responsible agentic architecture: Design systems where AI agents reason autonomously but consequential action is human-gated — with every decision writing an audit row recording what model decided what, on what evidence
• Governance-by-design: Embed deterministic classification/routing layers, human-in-the-loop oversight, model/data cards, and full audit trails into standard engineering practice — treating these as first-class architecture components, not afterthoughts
• Second-line posture: Ensure governance tooling supports independent review — maintaining separation between the teams that build and the functions that assess, with tool design reflecting this control
• Domain-specific requirements: Ensure AI systems handling financially material data (Finance), legally privileged documents (Legal), employee-sensitive information (HR), supplier-confidential data (Procurement), or audit evidence (Audit) meet the specific governance and evidential standards those domains require (e.g., SOX, legal privilege, chain-of-custody, employment law)
• Lifecycle governance: Own the ongoing governance obligations for live AI systems — periodic re-assessment, performance review against stated tolerances, change-impact assessment, and documented decision trails for model updates or retirement

4. Operational Resilience — For AI Systems

The second defining pillar. The role requires significant experience in operational resilience as a discipline — not just awareness, but hands-on engineering delivery.

• Engineer operational resilience into AI applications — circuit breakers, provider fallbacks, graceful degradation, fail-safe defaults, and dependency-aware architecture so essential functions survive outages
• Design for failure: assume model providers, data sources, and integration points will fail, and ensure user-facing services degrade safely rather than catastrophically
• Ensure AI systems are mapped against the organisation's important business services framework — with defined impact tolerances, recovery objectives, and tested failover paths

5. Engineering Leadership & Team Development

• Lead a multi-disciplinary engineering team comprising software, ML, data, and platform engineers — recruiting, developing, and retaining strong technical talent
• Set engineering culture and standards for code quality, testing, documentation, peer review, and production readiness
• Develop senior technical contributors and engineering leads — building depth and succession within the team
• Mentor and enable non-technical colleagues across enabling functions to ship their own AI applications, building AI literacy across HR, Finance, Procurement, Legal, Audit, and Compliance
• Manage capacity and team topology — making deliberate choices about structure, specialisation, and balance across discovery, delivery, and sustainment

6. Delivery & Production Excellence

• Drive end-to-end engineering delivery from architecture through to production deployment, scaling, monitoring, and lifecycle management
• Partner with enterprise technology teams to leverage shared platforms, infrastructure, and services — ensuring solutions are built on common foundations and contribute back to enterprise capability
• Establish and enforce standards for model serving, data pipelines, API design, integration patterns, security, and observability
• Drive MLOps maturity including CI/CD, automated testing, performance monitoring, incident management, and capacity planning
• Own the data engineering approach for AI across enabling functions — pipelines, feature stores, and data products designed for quality, governance, and reuse
• Establish data quality, lineage, and governance standards appropriate for financially material data, legally privileged documents, employee PII, and personally identifiable information subject to applicable data protection laws

7. Model Quality, Drift & Assurance

• Establish model-drift and bias monitoring across the portfolio — measuring divergence across runs and generating targeted refinement recommendations
• Design and operate AI red-team capability to probe and refine model outputs before release and on an ongoing basis
• Build automated fairness testing, explainability pipelines, and decision audit trails embedded in standard workflows
• Own technical risk management across the portfolio including model degradation, data quality drift, dependency risk, and integration risk
• Ensure model assurance reporting meets the expectations of internal audit, technology governance, and external regulators

8. Scaling Through Partnership

• Work in close partnership with enterprise technology to align on shared infrastructure, platform services, tooling, and engineering standards
• Identify and scale opportunities across enabling functions — recognising where a solution proven in one function (e.g., contract intelligence in Legal) can be adapted and deployed in another (e.g., supplier risk in Procurement)
• Engage external technology partners strategically — co-developing or integrating capabilities where they accelerate delivery or provide specialist functionality
• Shape joint programmes with partners, defining technical scope, integration architecture, and quality standards — ensuring partnerships deliver production-grade outcomes
• Build a scaling model that leverages shared services, external partners, and cross-functional reuse to multiply impact without proportional headcount growth

9. Stakeholder Partnership & Adoption

• Serve as a credible technical partner to functional leaders across HR, Finance, Procurement, Legal, Audit, and Compliance — providing counsel on AI opportunity, risk, regulatory posture, and investment
• Work as an internal consultant: run discovery with each business unit, surface real problems, frame concrete use cases with measurable success criteria, and drive adoption through continuous UX iteration and direct user-feedback loops
• Translate between engineering complexity and business strategy — communicating trade-offs, timelines, and constraints so leaders can make informed decisions
• Measure success by real adoption and operational impact, not by demonstrations — building KPI/KRI frameworks giving leaders visibility into usage, quality, and value

10. Enterprise AI Integration

• Represent enabling functions within enterprise AI leadership forums and architecture governance, ensuring alignment on standards, platform strategy, and shared services
• Contribute to enterprise AI governance including architecture review boards, AI policy development, and standards evolution
• Drive reuse and knowledge sharing — identifying where governance and resilience patterns developed for enabling functions can accelerate delivery elsewhere
• Advocate for enabling function requirements within the broader enterprise technology agenda, ensuring platform priorities reflect the needs of regulated, process-critical functions

Date Posted

Closing Date

AstraZeneca embraces diversity and equality of opportunity.  We are committed to building an inclusive and diverse team representing all backgrounds, with as wide a range of perspectives as possible, and harnessing industry-leading skills.  We believe that the more inclusive we are, the better our work will be.  We welcome and consider applications to join our team from all qualified candidates, regardless of their characteristics.  We comply with all applicable laws and regulations on non-discrimination in employment (and recruitment), as well as work authorization and employment eligibility verification requirements.