Director, Clinical AI Research

AstraZeneca • Boston, MA

13d • Hybrid

About The Position

AISI AI Strategy & Innovation (AISI) sits at the centre of AstraZeneca's R&D AI transformation. Our remit is to build, buy and deliver the AI models and agents that change pipeline outcomes, across discovery, translational science, biomarkers and clinical development. AstraZeneca is building a world-class AI capability for Clinical Development within AISI to accelerate the design, conduct, and analysis of clinical trials across our Oncology and BioPharmaceuticals pipeline. We are hiring an AI Engineer to build, post-train, evaluate, and deploy the next generation AI and agentic systems that power AstraZeneca's trials. AI for clinical development is a field in motion. Foundation models, multimodal learning, agentic systems, post-training methods, and evaluation science evolve rapidly and the regulatory and methodological frameworks around them are evolving in parallel. You'll identify and implement new methods as the field redefines what's possible. Comfort with ambiguity, an instinct to learn, and a bias for shipping are core to the role. The AI for Clinical Development function is being built from the ground up. As a lead engineer, you'll have unusual influence on the technical decisions, the eval and infrastructure choices, and the shape of the work. You’ll also have direct exposure to the regulators, scientific consortia, and external partners writing the rules of the road for AI in clinical evidence. We're looking for someone who can take a clinical problem from sketch, through model and agent design, through eval, and back through the feedback loop, and who understands that every model that ships as something that will eventually touch a trial, and a patient.

Requirements

• PhD in Computer Science, Machine Learning, Computational Linguistics, Biomedical Informatics, or a closely related computational discipline; or MSc with equivalent industry experience with large language models, agentic systems, and their evaluation.
• Minimum 3 years demonstrated, hands-on expertise in modern LLMs, multimodal models, and agentic systems, including end-to-end post-training, tool-use and agent harness design, and inference / serving in production.
• Excellent software engineering skills: Python, PyTorch, Hugging Face, cloud platforms (e.g., AWS, Azure, GCP), and modern LLM tooling.
• Track record of building and shipping AI systems that have been deployed in real clinical, healthcare, or other high-stakes domains.
• Expertise in sandboxing, containerization, VM infrastructure, and/or distributed systems.
• Strong experience constructing and using domain-specific evaluation environments and benchmarks for clinical or biomedical AI.
• Experience with AI safety and responsible AI methods applied to clinical AI.
• First- or co-first-author publications at top ML venues (NeurIPS, ICML, ICLR, CL).
• Excellent written and verbal communication, and ability to translate technical findings for clinical, regulatory, and executive audiences.
• Eager to work across the ML / clinical / operations / regulatory boundary.

Nice To Haves

• Direct experience with clinical NLP and clinical information extraction.
• Experience constructing hard, useful clinical benchmarks relevant to modern AI methods.
• Open-source contributions, leadership on widely-used clinical AI artifacts (datasets, benchmarks, models, evaluation tooling), and/or workshop and standards-body participation (e.g., Clinical NLP Workshop, ML4H, GenAI4Health, ML / clinical reporting standards).
• Prior engagement with FDA, NCI, NIH, or other regulators or public-health bodies on AI methodology, evaluation standards, or AI safety.
• Mentorship of junior engineers/computer scientists.

Responsibilities

• Build the next generation of agentic environments for AstraZeneca's clinical development workflows.
• Lead the team's frontier AI safety and oversight work.
• Collaborate across research and infrastructure teams to develop and ship environments.
• Debug and iterate rapidly across research AI/ML stacks.
• Contribute to research culture through technical discussions and collaborative problem-solving across AI and clinical teams.
• Train/post-train models for clinical alignment.
• Design agent harnesses, tool-use scaffolding, and multi-step orchestration for high-leverage clinical-development workflows.
• Build domain-specific evaluation environments and benchmarks for frontier clinical AI with end-to-end tracing, automated trace-level scoring, and drift/oversight detection.
• Partner with clinical development domain experts including clinicians, biostatisticians, translational scientists, clinical operations, and regulatory to scope problems, ship prototypes, and iterate on real-world feedback.
• Contribute to the team's reusable infrastructure for evaluation, monitoring, and oversight of AI in clinical development.
• Publish at top ML venues and top clinical and biomedical journals to maintain scientific credibility and external recruiting pull.
• Mentor more junior engineers and scientists, and represent AstraZeneca externally in industry forums and scientific meetings.

Benefits

• qualified retirement program [401(k) plan]
• paid vacation and holidays
• paid leaves
• health benefits including medical, prescription drug, dental, and vision coverage
• short-term incentive bonus opportunity
• equity-based long-term incentive program
• retirement contribution
• commission payment eligibility

What This Job Offers

• Job Type: Full-time
• Career Level: Director
• Education Level: Ph.D. or professional degree