Senior Director MMAI & Outcome Prediction – AI for Precision Health

We're building a connected, end-to-end Enterprise AI engine - uniting data foundations, AI technology, process reinvention, and business-facing AI to accelerate results across the whole value chain. Success depends on being exceptional connectors : you'll actively leverage existing capabilities, celebrate and promote reuse, export breakthrough ideas across geographies and functions, and obsess over scaling impact rather than building in isolation. If you thrive in high-collaboration environments where your role is to turn complex, cross-functional problems into reusable, enterprise-wide capabilities - and where the measure of success is adoption and scale, not just innovation - you'll have the platform (and sponsorship) to make it real. As Senior Director, Multimodal AI & Outcome Prediction within Enterprise AI – AI to Transform Care at AstraZeneca, you will lead the scientific translation of multimodal artificial intelligence and foundation model advances into clinically actionable capabilities across Oncology and BioPharma. Working in close collaboration with Enterprise AI, R&D teams, and AI for Science Innovation (AISI), you will drive the development, reinforcement, and validation of multimodal predictive and diagnostic systems integrating radiology, digital pathology, multi-omics (genomics, transcriptomics, proteomics), molecular diagnostics, clinical trial datasets, real-world electronic health records and claims, and longitudinal patient signals including digital biomarkers. Your work will enable the discovery and validation of AI-derived multimodal biomarkers and computational disease taxonomies that improve early diagnosis, refine disease stratification, support companion and AI-enabled diagnostic strategies, identify comorbidities, and guide treatment selection and responder identification. By applying advanced representation learning, outcome modelling, and survival analytics, you will translate multimodal intelligence into clinical development impact through trial enrichment, patient identification, endpoint optimisation, and deeper reanalysis of clinical trial data. In parallel, you will help reinforce foundation models using AstraZeneca’s multimodal trial and real-world datasets, creating continuous learning systems that connect discovery, development, diagnostics, and real-world outcomes across the product lifecycle. The role will also establish enterprise scientific standards for multimodal AI, including validation frameworks, cross-site robustness, regulatory-grade evidence generation, and performance monitoring, ensuring that AI-enabled diagnostic and predictive models can be trusted, scaled, and deployed to improve patient outcomes and accelerate precision medicine across the portfolio. Key Mission 1. Scientific Leadership in Multimodal AI and Computational Diagnostics Act as the enterprise scientific authority for multimodal AI applied to Oncology and BioPharma. Define and drive the scientific agenda for predictive modelling and computational diagnostics by developing advanced multimodal methodologies integrating imaging, molecular diagnostics, omics data, clinical trial datasets, digital biomarkers, and real-world evidence. Champion methodological excellence in multimodal representation learning, computational imaging, omics integration, disease trajectory modelling, and survival prediction. Ensure the scientific rigor, reproducibility, and robustness of AI models used to derive predictive biomarkers, diagnostic intelligence, and patient stratification strategies. 2. Advance Diagnostic Innovation and Computational Disease Stratification Lead the development of AI-enabled diagnostic frameworks that combine imaging phenotypes, molecular signatures, and clinical data to identify disease states earlier and refine biological disease taxonomy. Drive the discovery and validation of multimodal biomarkers that support early diagnosis, disease subtype classification, and treatment selection . Contribute to the development of companion diagnostics and AI-enabled diagnostic strategies aligned with precision medicine and regulatory requirements, enabling improved patient identification and clinical decision support. 3. Transform Clinical Development Through Predictive Intelligence Apply multimodal AI methodologies to transform clinical development strategies by improving patient identification, trial enrichment, responder prediction, and endpoint optimisation. Lead advanced reanalysis of clinical trial datasets to uncover responder subgroups, identify predictive and prognostic biomarkers, and refine patient selection strategies. Use advanced modelling approaches such as causal inference, treatment effect estimation, and dynamic outcome prediction to strengthen development decisions and maximise asset differentiation across the portfolio. 4. Reinforce Foundation Models with Clinical and Real-World Data Partner closely with internal AI research teams to translate advances in foundation models into practical biomedical applications. Design reinforcement strategies that leverage AstraZeneca’s clinical trial datasets, real-world healthcare data, and multimodal biological signals to improve model generalisability and predictive power. Develop reusable multimodal representations that capture disease biology across datasets and therapeutic areas, enabling scalable predictive modelling capabilities across the organisation. 5. Integrate Clinical Trials and Real-World Evidence into Continuous Learning Systems Establish predictive modelling frameworks that integrate clinical trial data with real-world evidence to extend insights beyond controlled trial environments. Develop continuous learning systems capable of incorporating longitudinal patient outcomes from electronic health records, claims data, and diagnostic platforms. Enable post-launch monitoring of treatment outcomes and reinforcement of predictive models through real-world evidence, creating feedback loops that strengthen both development and care pathway strategies. 6. Establish Enterprise Standards for Multimodal AI Validation and Governance <