PhD Position in Medical AI and Foundation Models

We, the research group led by Michael Moor, are looking to fill a PhD position focusing on medical AI and medical foundation models. We are looking for an ambitious candidate eager to conduct groundbreaking research in medical AI. Our goal is to advance the capability, reliability, and safety of Multimodal Large Language Models (MLLMs), Vision Language Models (VLMs), and text-only LLMs across various medical and biomedical contexts. The successful candidate will join a cutting-edge research group working on projects that include but are not limited to clinical retrieval systems, representation learning models, early warning systems, and interactive multimodal chat applications.

• Design, develop (pre-train, SFT, and preference-align) and benchmark new medical foundation models across different data modalities
• Conduct research on retrieval augmentation strategies and its integration into explainable clinical decision support systems
• Develop novel methods to address model hallucinations, and the reliability problem of LLMs and MLLMs, specialized for high-stakes domains such as healthcare
• Develop new strategies to encapsulate and organize medical knowledge and to make it accessible to medical AI systems
• Publish research findings in top-tier machine learning conferences as well as top-tier biomedical and general scientific journals (depening on project)
• Participate in academic collaborations within ETH Zurich, across Swiss institutions as well as with leading international collaboration partners

• Master’s degree in Computer Science, Artificial Intelligence, or a relevant related field (e.g. Mathematics, Electrical Engineering etc) OR
• Master’s degree in Medicine (or MD), plus intermediate-to-advanced Python coding skills (ideally some previous ML / AI experience in Python)
• Strong programming skills in Python and familiarity with machine learning and NLP frameworks such as PyTorch, huggingface-related frameworks, peft, deepspeed etc
• For computational candidates: prior experience with large-scale ML training (multi-node training, distributed training strategies, training of >10B trainable parameters) is a big plus
• For all candidates: prior experience in machine learning, especially in LLMs, VLMs, or related areas is highly welcome
• Strong computational “housekeeping skills” are a plus: batched job submission, bash scripting, code version control, data version control, etc
• Demonstrated ability to work independently as well as part of a team
• Excellent communication skills in English, both written and spoken
• Prior ML engineering / research experience in AI industry is welcome
• Prior top-tier publications are not required, but are a plus

The position is located in the Department of Biosystems Science and Engineering (D-BSSE) of the ETH Zurich in Basel, Switzerland. The D-BSSE is a highly interdisciplinary department – specializing in biological engineering, systems and synthetic biology, bioinformatics and data science, and engineering sciences related to microfluidics and microfabrication. The D-BSSE is centrally located within a biomedical research hub with close links to top academic institutions (e.g., Friedrich Miescher Institute (FMI), Biozentrum and University of Basel) as well as major biotechnology and pharmaceutical companies (e.g., Novartis, Roche, Bayer, and Lonza). The ETH Zurich is a global leader in science and technology and consistently ranks as one of the top universities in the world. Basel, Switzerland is an international city on the border with France and Germany – nested between the Swiss alps and the black forest. The city provides easy access to arts and culture, nature and adventure, and short commutes via train/plain/automobile to anywhere in Europe.