University of Cambridge

Job Opportunities


PhD Studentship - Combining machine learning with Physics to design soft materials (Fixed Term)

Applications are invited for a fully funded PhD studentship (Home/EU/International) in Dr Alpha Lee's group ( on designing soft materials by combining physics with machine learning. The studentship has NO nationality restrictions.

The PhD programme associated with this studentship would start in October 2018. The award covers tuition fees (for UK/EU/international students) and provides a tax-free stipend of £14,553 p.a. (index linked). The University has a responsibility to ensure that all employees are eligible to live and work in the UK.

Fixed-term: The funds for this post are available for 3.5 years in the first instance. The student must complete the programme in 42 months.

The successful candidate should have a good first degree and a Masters pass in a quantitative field (e.g. physics, chemistry, mathematics, computer science, statistics) that is relevant to the projects. The successful candidate must be highly motivated, capable of performing independent research and have excellent communication skills with the ability to work collaboratively.

Recent technological advances have made high throughput simulations and experiments of soft materials possible. However, the analysis of voluminous and high-dimensional data demands an innovative and novel approach that integrates data into physical theories. Our group is interested in closing the molecules-to-materials gap by applying physics-based machine learning techniques to organic molecules. We seek an enthusiastic PhD student to work on one of our two primary research areas: designing organic electrolytes and computational drug discovery.

Renewable energy sources such as solar and wind are intermittent. Overcoming intermittency is the key to transiting from fossil fuel to renewables. This requires energy storage devices that have both large energy density and power density, yet existing devices either have a large energy density or power density but not both. Electrical double layer supercapacitors, which store energy by accumulating ions near oppositely charged electrodes, are promising devices that could resolve this capacitance-power dilemma. This project will develop computational methodologies to screen electrolyte-solvent-electrode combinations by using machine learning to interpret and design molecular simulation and experiments. We will parameterize molecular-scale models and continuum theories, using machine learning to bridge between molecular structure and condensed phase behaviour. We will also derive Bayesian statistical methods to decide which simulation methodology to use based on a trade off between computational time and predicted uncertainty. This project builds on analytical models that we have derived for concentrated electrolytes and supercapacitors (Physical Review Letters, 119, 26002 (2017); Physical Review X, 6, 21034 (2016); Physical Review Letters, 115, 106101 (2015).

The second project investigates machine learning methods to infer representations of organic molecules to accelerate preclinical drug discovery. We will develop deep learning methods to make inference on large (and potentially noisy and incomplete) pharmacological datasets, refining traditional notions of chemical similarity and pharmacophores to make them statistically powerful. We will also develop scalable Bayesian methods to estimate the uncertainty of the algorithms, and drive experiments using machine learning models. Another dimension that we are interested in are methods that predict the outcome of organic reactions by combining quantum chemical insights with data from high throughput synthesis studies. This project builds on our work on data-driven models for predicting protein-ligand affinity (Proceedings of the National Academy of Sciences, 113, 13564 (2016)) and designing high throughput experiments that combine coarse and fine data (Physical Review Letters, accepted (2017) [arXiv:1702.06001]) .

Interested candidates are encouraged to make informal enquiries by contacting Dr. Alpha Lee (

To make an application, follow the procedure outlined on the University website, selecting the course PhD in Physics and making sure to mention the name of Dr Alpha Lee and the Theory of Condensed Matter group. Awards may also be made to supplement part-support from other sources, and candidates are encouraged to express their interest for other available awards in the application form in addition to this Studentship, and also apply to the Winton Scholar programme (

It is IMPORTANT that, when submitting the application, Dr Alpha Lee is also notified through an e-mail to

Please quote reference KA13863 on your application and in any correspondence about this vacancy.

The University values diversity and is committed to equality of opportunity.