AI Impact & Applied Innovation Specialist

ASTRAZENECA UK LIMITED; Cambridge; On-site; GBP 110,000 - 170,000; Full time

Job summary

AstraZeneca UK Limited is seeking a Senior AI Impact & Applied Innovation Specialist to translate frontier AI into practical oncology R&D insights. You will lead cross-department projects, map high-value AI questions, and guide evaluation frameworks to ensure responsible deployment in drug discovery and development.

Lead scientific translation from frontier AI to actionable decision-making, coordinating expertise across AI science, oncology R&D, and Biopharma.

Qualifications

• Advanced degree in quantitative biology or computational discipline is required.
• Experience applying AI/ML to clinical, genomic, transcriptomic, or biomedical data in drug development.
• Strong scientific literacy across oncology drug development pipeline.
• Ability to frame rigorous scientific questions and translate to tractable AI problems.
• Experience evaluating AI model quality in clinical contexts.
• Track record of working across multidisciplinary teams and communicating to senior audiences.

Responsibilities

• Maintain an interactive framework of AZ AI capabilities and frontier models.
• Identify high-priority scientific questions where frontier AI adds value.
• Design evaluation approaches for generative and agentic systems and run red-teaming.
• Translate hypotheses into time-bound projects with clear success criteria.
• Communicate AI science priorities and recommendations to senior stakeholders.

Job description

Are you ready to translate AstraZeneca's frontier AI capabilities into concrete impact for the organisation? The Applied AI Research team, part of the AI Science Innovation (AISI) group within Enterprise AI (EAI), sits at the intersection of frontier AI research and oncology R&D, with a mission to ensure that advances in foundation models, agentic systems, and biological AI find their way to the problems where they will have the greatest impact on how we discover and develop medicines. As Senior AI Impact & Applied Innovation Specialist, Applied AI Research, you will be involved in both sides of that translation: identifying where frontier AI can address fundamental drug discovery questions for AstraZeneca, and orchestrating focused cross-departmental projects to answer them. You will maintain a structured view of AZ's highest-value AI science problems - mapping which questions most need answers, which frameworks are now mature enough to address them, and where AZ's unique data, biology, and portfolio create irreplaceable scientific advantage. You will also act as the scientific lead and orchestrator for delivering against those problems, assembling the right combination of expertise and resources and holding the scientific hypothesis through to a result that changes a decision. This is a role for someone familiar with key questions across the drug discovery pipeline, who can attend a frontier AI seminar and immediately see which ideas are applicable to areas from target identification to trial design, and who has the scientific conviction to advocate for a particular problem in the face of competing priorities. If you thrive at the intersection of machine learning and translational oncology and want to leave a direct mark on the scientific questions that shape AstraZeneca's pipeline, this role is for you.

Typical Accountabilities AI Landscape & Opportunity Mapping:

• Maintain an interactive framework providing a structured, dynamic inventory of AZ's internal and external AI capabilities, including frontier models and platforms.
• Identify high-priority scientific questions-such as those in target discovery, combination therapy, tumor mapping, and trial design-where frontier AI can deliver significant value. Distinguish opportunities unique to AstraZeneca from those addressable with generic tools or public data.

Scientific Evaluation Standards:

• Apply scientific expertise and evaluation standards to AI-driven frameworks designed to maintain a live, structured record of R&D goals and decision-making.
• Design and apply evaluation approaches for generative and agentic systems, including harnesses, benchmarks, trajectory-level metrics and validated LLM-as-judge methods, and run structured red-teaming (adversarial testing, failure-mode taxonomies, pre-release and continuous monitoring).
• Advise on and test where AI agents can responsibly automate steps in scientific, clinical and operational research processes, defining oversight points, acceptance criteria and comparison with the manual baseline.

Forward-Deployed Embedding & Delivery:

• Translate scientific hypotheses into well-specified, time-bounded projects with pre-defined success criteria, with a clear path to integration and deployment; orchestrate the delivery of these projects
• Establish frameworks for quantifying impacts of AI projects and measuring impact for the business and implement these for active cross-functional projects.

Stakeholder Communication & Leadership:

• Communicate AI science priorities, programmed status, and strategic recommendations to senior stakeholders across EAI, Oncology R&D and Biopharma R&D; act as the scientific bridge between AISI strategy and execution-layer teams, ensuring that frontier insight translates into actionable work and that delivery blockers are visible to leadership.

Qualifications

• Advanced degree (PhD preferred; Master's with substantial relevant experience considered) in a quantitative biological or computational discipline: computational biology, bioinformatics, machine learning, statistics, or a closely related field.
• Demonstrated experience applying AI/ML methods to clinical, genomic, transcriptomic, or other biomedical data in a drug development or translational research context.
• Strong scientific literacy across the oncology drug development pipeline, from target identification and preclinical biology through biomarker development, clinical trial design, and Phase 1/2 decision-making.
• Familiarity with the current landscape of AI in drug development, including foundation models for biological and clinical data, perturbation prediction, digital pathology, and agentic research systems.
• Demonstrated ability to frame rigorous scientific or clinical questions and translate them into tractable AI problems with well-specified success criteria.
• Experience evaluating the quality of AI models in a scientific or clinical context, including understanding of validation standards, confounding, shortcut learning, and the distinction between prognostic and predictive signals.
• Track record of working effectively across multidisciplinary teams, including both technical AI/ML specialists and experimental or clinical scientists.
• Excellent written and oral scientific communication, including the ability to present technical material to non-technical senior audiences and to distil complex landscapes into actionable recommendations.
• Comfort with ambiguity and a bias for action; able to make well-reasoned prioritization decisions under uncertainty and to advocate for them.

Desirable Requirements:

• Direct experience with generative AI and agentic systems applied to life sciences, particularly for clinical insights.
• Direct experience with multimodal foundation models, particularly for computational pathology, for response prediction, subtyping and other common research tasks.
• Direct experience with multimodal data integration, for example combining clinical, genomic, transcriptomic, imaging, and/or pathology data.
• Familiarity with CRISPR functional genomics, perturbation prediction, or virtual cell approaches and their application to target discovery or combination prioritization.
• Familiarity with AI benchmarking methodologies and the challenges of evaluating foundation models on clinically meaningful tasks.
• Experience with AI prioritization or portfolio management methodologies - including problem identification, value assessment, and structured decision frameworks.
• A record of scientific contribution (e.g., peer-reviewed publications), ideally including AI/ML or clinical AI venues, that establishes credibility with technical, clinical, and scientific audiences