About the Company
Boehringer Ingelheim is a biopharmaceutical company active in human and animal health. It develops innovative therapies in areas of high unmet medical need and invests in research and development.
Responsibilities
•Contribute engineering perspectives to architectural design discussions to support efficient training and inference, scalability, and production readiness
•Implement biomedical foundation model components, including training code, data loaders, tokenisers, inference logic, and fine-tuning interfaces, under the guidance of senior ML engineers
•Collaborate with AI scientists to translate validated research prototypes into robust, production-quality model artefacts and contribute to benchmarking and performance evaluation
•Optimise validated models and inference pipelines using techniques such as quantisation, distillation, and pruning to meet production efficiency and latency requirements without compromising performance
•Write clean, well-tested, and well-documented code in line with the team's engineering standards
•Support model handovers to MLOps engineers and document capabilities, known limitations, failure modes, and retraining criteria
•Stay current with advances in ML engineering, distributed training, and biomedical AI tooling
Requirements
•Postgraduate degree in Machine Learning, Computer Science, Computational Biology, or a related technical field
•Hands-on experience with deep learning and foundation model implementations, such as transformers, pre-training, and fine-tuning
•Some experience contributing to production-quality model artefacts and a growing understanding of moving research prototypes to reliable deployment
•Experience collaborating with researchers during implementation
•Proficiency in Python and deep learning frameworks such as PyTorch
•Strong software engineering fundamentals, including clean, testable, documented, and maintainable code, version control, and code reviews
•Working knowledge of distributed training frameworks such as PyTorch Distributed, DeepSpeed, FSDP, or Ray Train
•Exposure to inference optimisation techniques such as quantisation, distillation, and pruning
Nice to Have
•PhD or equivalent industry experience with an MSc
•Experience with biomedical data modalities such as genomics, multi-omics, clinical or imaging data in an ML context
•Publications
•Contributions to open-source ML projects or tooling
Conditions
•This is a hybrid role with 3 days a week in the office
•Recognised as a Top Employer in the UK
•Access to health and wellbeing programmes and groups
