ML Engineer

In this role you will contribute to building production-ready biomedical foundation models as part of the AI Systems team within the AI Accelerator. You will help translate validated research into scalable implementations and optimize model training and inference for real-world health outcomes. You’ll work closely with AI scientists to ensure engineering rigor and maintainability across the stack. You’ll operate at the intersection of biology and AI, shaping how models inform decisions on indications, patient populations and targets.

• hybrid role with ~3 days in the office
• Top Employer in the UK recognition
• strong people practices and supportive HR policies
• opportunity to work at frontier of biomedical AI
• cross-disciplinary collaboration
• focus on impact on human health

• Provide engineering input on architectural design to ensure scalable, efficient training and inference
• Implement core model components (training code, data loaders, tokenisers, inference logic, fine-tuning interfaces) to a high standard
• Collaborate with AI scientists to convert research prototypes into robust, production-quality artifacts and support benchmarking
• Optimize validated models and inference pipelines using techniques like quantisation, distillation and pruning to meet latency and efficiency targets
• Write clean, well-tested, well-documented code and adhere to team engineering standards
• Support handover to MLOps engineers with documentation on capabilities, limitations, failure modes and retraining criteria
• Stay updated with advances in ML engineering, distributed training and biomedical AI tooling

• Postgraduate degree in ML, CS, Computational Biology or related field; PhD preferred or MSc with equivalent industry experience
• Hands-on experience with deep learning and foundation models (transformers, pre-training, fine-tuning)
• Experience delivering production-quality model artifacts and moving from research to deployment
• Experience collaborating with researchers during implementation
• Proficiency in Python and PyTorch
• Strong software engineering fundamentals (clean, tested, documented code; version control; code reviews)
• Experience with distributed training frameworks (PyTorch Distributed, DeepSpeed, FSDP, Ray Train)
• Exposure to model optimization for inference (quantisation, distillation, pruning)

• Cross-functional collaboration
• Attention to engineering craft and documentation
• Ability to translate research into production-ready solutions
• Python
• PyTorch
• Transformers