Senior Scientist, Clinical Agentic AI

About the role

AstraZeneca is hiring a Senior Scientist, Clinical Agentic AI for an in-office role in Barcelona, Spain, with a minimum of three days per week on site; remote or travel flexibility is not available. The role sits in a hands-on clinical AI group within the Enterprise AI engine, working with clinical and R&D partners to turn high-value clinical workflows into systems that reason over documents and data, use enterprise tools and expose their evidence so clinicians can trust and adopt them.

Key responsibilities

• Design and deliver LLM-driven, agentic workflows for trial planning, conduct, monitoring, review and reporting that improve speed, quality and evidence transparency.
• Write production-quality Python and SQL, build services, data pipelines, retrieval systems, tool integrations, structured-output components and automated tests meeting reliability, latency and cost targets.
• Integrate structured and unstructured clinical sources with provenance, access controls and data contracts, and implement verification, citations, uncertainty handling and human-in-the-loop review so systems fail safely.
• Build reproducible experiments and evaluation datasets, select meaningful metrics, perform error analysis and compare models, retrieval strategies and orchestration patterns, turning findings into concrete improvements with release evidence.
• Co-create solutions with CRAs, clinical scientists, medical monitors and operations teams, iterating through rapid prototyping and user feedback, and contribute to design reviews, code reviews, documentation, risk assessments and reusable components.
• Track advances in agentic AI, clinical NLP, multimodal models, evaluation and responsible AI, reproduce promising methods, present results and contribute to internal standards and external scientific engagement.

Required profile

• PhD in Computer Science, Machine Learning, Biomedical Informatics, Computational Biology, Statistics or a related quantitative field; or a master's degree with equivalent applied research and engineering experience.
• 1 to 3 years of post-PhD (or master's-level equivalent) experience building applied ML, NLP, generative-AI or agentic systems beyond notebooks or demonstrations.
• Understanding of modern language-model techniques, retrieval-augmented generation, embeddings, tool use, structured generation and evaluation.
• Ability to design controlled experiments, select metrics, perform error analysis and communicate uncertainty, and to trace outputs back to their sources in complex, heterogeneous data.
• Curiosity about clinical development and clear written and verbal communication with product, engineering, clinical and quality colleagues.

Required skills

• Strong hands-on Python and working knowledge of SQL, software testing, version control, APIs and reproducible development.
• Desirable: clinical NLP, biomedical language models, multimodal data, trial protocols, EHR, EDC, CTMS, eTMF, safety data, agent frameworks, workflow engines, cloud, containers, CI/CD, model serving, observability, MLOps/LLMOps, knowledge graphs, ontologies, document intelligence, GCP and GxP.