Senior Director, Machine Learning & AI (BPD)

Role purpose

AstraZenecas bold ambition is to be a pioneer in science lead in our disease areas and transform patient outcomes and by 2030 to deliver 20 new medicines and industryleading growth. Biologics are central to that ambition and Biopharmaceutical Development (BPD) is the R&D function that turns biologic candidates into medicines.BPD develops the cell lines bioprocesses formulationsdevicesand analytical methodsneeded to advance biologic medicines through clinical development and approvalwhere they can improve the lives ofpatients. As the portfolio grows in scale and complexity BPD is increasingly adopting a PredictFirst CMC approach: FAIR data at source greater use of modelling and digital twins and AI-enabled tools that help scientists find knowledge make decisions and create regulatory content more efficiently.

The Senior Director Machine Learning & AI leads the ML & AI team within BPD: a multidisciplinary group of specialists spanning data science AI and data engineering and applied machine learning research. The role is accountable for translating BPDsPredict Firstambition into a coherent AI strategy and portfolio roadmap that transforms emerging technologies and promising ideas into trusted scalable capabilities that deliver measurable scientific and business value. The Director defines the ML & AI strategy for BPD owns delivery of the AI portfolio within the digital transformation roadmap and serves as BPDs senior technical interface with Enterprise AI and R&D IT. The roleis responsible forestablishinga framework that rapidly tests anddemonstratesvalue through proof-of-concepts (PoCs) accelerates adoption through iterative delivery and enables the scaling of successful AI solutions across BPD.

In addition the Director partners closely with Robotics & Automation Informatics Digital Transformation Enterprise AI and R&D IT teams to identify opportunities where ML & AI can enhance scientific operational and business outcomes and to integrate AI capabilities into products platforms and workflows across BPD(e.g.PhysicalAI). The role provides strategic leadership on the data foundations required to enable AI at scale including data architecture governance engineering and platform capabilities ensuring that high-quality accessible and trusted data can support advanced analytics machine learning and AI solutions across the enterprise.

Success in this role requires a balance of strategic leadership and technical credibility. The Director will shape investment decisionsbuild organisational capability drive adoption across BPD influence senior stakeholders across BPD and the enterprise and provide the technical judgement needed to guide delivery and manage risk.

Key accountabilities

Strategy and portfolio

• Define andmaintainBPDs multi-year ML&AI strategy aligned withaPredictFirstCMCorganisation the BPD digital transformation roadmap and AZs AI30 ambitions.

• Be accountable for the BPD AI portfolio across the four pillars: AI Foundations & Platforms Knowledge Management Modelling & Digital Twins and Submission & Report Authoring.

• Set portfolio priorities across in-flight self-funded and proposed initiatives making clear evidence-based recommendations on when to build buy partner pause or stop.

Technical leadership

• Provide senior technical oversight of model strategy evaluation and deployment across predictive ML mechanistic and hybrid models protein sequence and structure models knowledge graphsRAGand agentic architectures.

• Set practical engineering standards for the team including reproducibility model risk managementMLOps evaluationframeworksand human-in-the-loop approaches forGxP-adjacent use cases.

• Chair or lead technical review of the highest-risk or highest-value deliverables ensuring decisions are wellevidencedand risks are visible to the right governance forums.

Team leadership

• Leadanddevelopahigh-performingML&AI team of data scientists and AI/data engineers growing capability and reach through permanent hires secondments PDRAs and vendor partnerships.

• Create the operating modelownershipand delivery discipline needed for a small specialist team to have enterprise-level impact.

• Support AI training and culture change across BPD helping scientists use AI well rather than simplyuseit more.

Crossfunctional delivery

• Workwith modelling/AIdigitalisationand robotics transformation leads toaligninvestment dependencies and delivery plans across AIdataand automation.

• Partner with R&DITso enterprise platforms meet BPDs scientific needs and BPD requirements are visible in strategic platform roadmaps.

• Serve as BPDs senior technical voice into Enterprise AI: adopt enterprise capability where it fits escalate gaps and shape shared offerings where BPD should not rebuild common capability

• Work closely with CMC Statistics Informatics & Software Engineering and Robotics & Automation Developmentcolleaguesso that ML&AI outputs sit on sound statistical software and laboratoryfoundations.BuildPhysical AI as an emerging BPD capability by partnering with Robotics & Automation Informatics Digital Transformation Enterprise AI and R&D IT to connect ML&AI models agents and decision-support tools with laboratory automation instrumentation and closed-loop experimental workflows.

Governancecomplianceand risk

• Ensure BPDs AI work aligns with AZ AI governance data governance informationsecurityandGxPexpectations as well as emerging external regulatory guidance on AI in CMC.

• Contribute to AZs regulatory advocacy on AI in CMC where BPDs experience is directly relevant (e.g.via the CMC Strategy Board and PMF AI in CMC Working Group).

• Be accountable forresponsible-AIpractice across the BPD portfolio including model documentation validation evidence bias and robustness testing and lifecycle management.

External innovation and partnerships

• Workwith the AI Partnershipsleadtobringuseful external thinking into BPD through academic collaborationsconsortiaand vendor evaluations.

• Represent BPD externally through selected publicationsconferencesand standards forums where this supports the strategy.

Stakeholder engagement

• Briefdigital transformation andBPDleadershipon progress valuetrade-offsand risk distinguishing clearly between proven capability activepilotsand speculative opportunities.

• Act as a trusted advisor to BPD functionalleaders onwhere AI can and cannot help them meet theirobjectives.

Qualifications and experience

Essential

• Advanced degree (MSc or PhD) in a quantitative discipline: computer science machine learning statistics applied mathematics physics computational biology chemical/biochemical engineering or a closely related field.PhDplus 7 yr of relevant experience. MSc plus 10 ye of experience.

• Track recordof leading ML&AI teams that deliver production capability not just prototypes in regulated or scientifically demanding environments.

• Deep current hands-on knowledge across the following: classical ML deep learning foundation or language models agentic systems digital twins knowledge graphsRAGandMLOps.

• Strong software engineering discipline; fluent in Python and modern ML tooling; comfortable working in cloud environments such as Azure or AWS includingcontainerisedworkloads and distributedcompute.

• Experience turning ambiguous scientific or business problems into shaped AI solutions including knowing when AI is not the right answer.

• Ability to influence senior stakeholders across scientifictechnicaland business functions and to make clear recommendations under uncertainty.

• Experience building durable partnerships with IT/platform teams externalvendorsand academic groups with clear commercial technical and delivery outcomes.

Desirable

• Domain understanding of biologics CMC bioprocess development formulation analytical development or regulatory submissions.

• Familiarity with FAIR data principles data product thinkingontologiesand knowledge graphs applied to scientific data.

• Experience withGxPadjacent AI model validation for regulated use or contribution to regulatory advocacy on AI/ML.

• Peer-reviewed publications orrecognizedexternal contributions in applied ML for life sciences.

What success looks like in the first 1218 months

• Measurable time saved on knowledge retrieval across BPD supported by an agent architecture and evaluation framework the team is confident to scale.

• At least one authoring pipeline moved from proof of concept into production use for a regulatory submission or comparability report.

• A working digital twin capability for aprioritisedunit operation with a defensible modelling strategy for the rest of the roadmap.

• An ML&AI team that is known inside BPD and beyond for highquality delivery clear technical judgement and honest communication about what AI can and cannot do.

• BPD requirements reflected in enterprise roadmaps deliverycommitmentsand platform investment decisions.