PhD position in Digital Chemistry

The Digital Chemistry Laboratory is led by Prof. Dr. Kjell Jorner at the Institute of Chemical and Bioengineering, within the Department of Chemistry and Applied Biosciences at ETH Zurich and associated with the ETH AI Center. We are an interdisciplinary group at the intersection of chemistry and computer science. Our mission is to accelerate chemical discovery using digital tools. We predict chemical reactivity and molecular properties using machine learning, artificial intelligence, computational chemistry, and cheminformatics. Our ultimate goal is the computer-aided design of molecules and catalysts.

Predicting the outcome of chemical reactions is important for optimizing chemical processes, synthetic route selection in the pharmaceutical industry and for discovery of new catalysts, among other things. In this project, we will create hybrid models that blend classical methods with expert-selected features and modern deep learning techniques, leveraging the strengths of both approaches. Quantitative reaction prediction involves important properties such as yields, selectivities and activation energies. Due to the low availability of experimental data, current prediction models are mostly based on expert-crafted features that are specific for each reaction type, and that are coupled with classical machine learning models such as decision tree ensembles and (regularized) linear regression. While deep learning models are more flexible, generally applicable, and perform better when more data is available, they fall short in the low-data regime that is the reality for most chemical reactions. In this project, we will use geometric deep learning models that exploit 3D molecular information, including graph neural networks and 3D convolutional neural networks, and combine them with traditional features. The project is highly interdisciplinary and will be carried out in close collaboration with groups in computer science and applied mathematics.

As a PhD student in our growing team, you will work with chemical reaction data to develop state-of-the art hybrids between classical descriptor approaches and modern geometric deep learning models to predict reaction properties such as activation energies, yields and selectivities. This will be done both with colleagues in the group and with our collaborators in computer science and applied mathematics. You will further contribute to the teaching activities of the group.

We are looking for a committed and motivated candidate that is excited to push the boundaries of research in digital chemistry. 

Essential experience, skills, and characteristics:

• A master’s degree in either chemistry, chemical engineering, computer science,  data science, physics, or related fields, or expectation of obtaining such a degree before February 2025
• English proficiency
• Self-motivation, ability to work independently and solution-oriented mentality.
• Interdisciplinary and collaborative mindset and desire to work with people from different disciplines and backgrounds
• Programming experience using languages such as Julia, Python, R, etc.

Desirable criteria:

• Experience with deep learning frameworks such as PyTorch, Jax, Tensorflow or Keras from research projects or thesis
• Knowledge of chemistry and chemical reactions from courses or projects
• Experience of applied machine learning from research projects or thesis
• Experience in geometric deep learning

A background in chemistry is not required, however a willingness to learn is crucial.

You will join a new, dynamic, and growing research group in the emerging field of Digital Chemistry at the highly motivating environment of ETH Zurich. We foster a modern and supportive group culture and value diversity, independence, and initiative. The position is embedded in an exciting and interdisciplinary research environment with connections to the ETH AI Center and the National Competence Center for Research, NCCR Catalysis, connecting the chemical sciences, digitalization, and sustainability.

A competitive salary is paid according to Rate 2 of the Doctoral student salary ladder.