Insights analyst

AstraZeneca•Gaithersburg, MD

1d•$72,120 - $108,181•Hybrid

About The Position

AstraZeneca is seeking an exceptional Scientist or Senior Scientist to join their AI/ML for Translational and Clinical Sciences team. This role involves working at the intersection of digital twins, foundation models, and multimodal clinical data. The successful candidate will design and build next-generation machine learning systems to predict clinical outcomes, discover novel biomarkers, and enable precision patient stratification across AstraZeneca's therapeutic areas. The work will involve building patient-level digital twins that integrate longitudinal clinical, imaging, genomic, proteomic, and real-world data, leveraging foundation models to reason across modalities and time. This role directly informs trial design, endpoint selection, and translational decision-making, accelerating the delivery of transformative therapies. It is a highly collaborative position interfacing with Data Science, Clinical Development, Translational Medicine, and Biometrics, with significant exposure to therapeutic area leadership.

Requirements

PhD in Computer Science, Machine Learning, Computational Biology, Biomedical Engineering, Biostatistics, Physics, or a closely related quantitative discipline OR an MS in a comparable discipline with equivalent applied research experience in AI/ML for healthcare or life sciences.
Demonstrable experience developing machine learning or deep learning models applied to clinical, biomedical, or omics data.
Strong proficiency in Python and modern ML frameworks (PyTorch, JAX, or TensorFlow), including experience with distributed training on GPU/TPU infrastructure.
Solid understanding of transformer architectures, self-supervised learning, and foundation model training or fine-tuning paradigms.
Experience working with longitudinal clinical data (EHR, clinical trials, registries) and familiarity with data standards such as OMOP, CDISC (SDTM/ADaM), FHIR, or DICOM.
Strong grounding in statistical inference, causal modelling, or survival analysis, and a rigorous approach to validation and generalisation.
Track record of scientific output through peer-reviewed publications, preprints, or open-source contributions.
Excellent written and verbal communication skills, with the ability to explain complex methods to non-technical stakeholders.
Scientist: PhD with 0–3 years of relevant post-PhD experience, or MS with 4+ years of relevant industry/research experience in applied ML for biomedical data.
Senior Scientist: PhD with 5+ years of relevant post-PhD experience, or MS with 8+ years of relevant experience, and a demonstrated record of leading end-to-end ML projects and influencing scientific or product strategy.

Nice To Haves

Experience building or contributing to digital twin, synthetic control, or mechanistic-ML hybrid models in a healthcare or life-sciences setting.
Familiarity with multimodal representation learning, including vision-language models, graph neural networks, or time-series transformers.
Prior work with multi-omics integration (genomics, transcriptomics, proteomics, single-cell) and pathway-informed modelling.
Experience deploying models in regulated environments (GxP, SaMD) and familiarity with model interpretability, uncertainty quantification, and fairness frameworks.
Exposure to cloud platforms (AWS, Azure, GCP), MLOps tooling, and reproducible research practices (containers, workflow managers, experiment tracking).
Experience collaborating within pharma R&D, clinical development, or academic medical centres.

Responsibilities

Design, train, and validate patient-level digital twin models that simulate disease trajectories and treatment response using longitudinal multimodal clinical data.
Contribute to the development, fine-tuning, and evaluation of foundation models (transformer-based, generative, and multimodal) tailored to clinical and biomedical data, including EHR, medical imaging, omics, and free-text clinical notes.
Build predictive and causal ML models for clinical endpoints, adverse events, disease progression, and treatment response, ensuring rigorous validation against prospective and external datasets.
Develop scalable pipelines and representation-learning approaches that unify structured clinical, genomic, transcriptomic, proteomic, imaging, and real-world evidence data.
Apply interpretable ML and causal inference methods to identify and validate novel prognostic and predictive biomarkers from clinical trial and real-world datasets.
Design ML-driven stratification strategies to support precision medicine hypotheses, enrichment trial designs, and companion diagnostic development.
Partner closely with clinicians, statisticians, translational scientists, bioinformaticians, and MLOps engineers to translate models into decision-grade tools embedded in R&D workflows.
Publish in top-tier venues (Nature Medicine, NeurIPS, ICML, Cell Patterns, Lancet Digital Health), represent AstraZeneca at external conferences, and contribute to strategic partnerships with academic and technology collaborators.
Ensure that models are developed in line with GxP, model risk management, fairness, privacy, and emerging regulatory guidance (FDA, EMA, MHRA) for AI/ML in drug development.
At the Senior Scientist level, shape scientific strategy, mentor junior scientists, lead cross-functional workstreams, and act as a technical authority in digital twin and foundation model methodology across the portfolio.