Research Areas

Autonomous Science

One primary research interest is in delivering autonomous drug discovery systems, merging synthetic organic chemistry, laboratory automation and machine learning. The aim and innovation lie in automating the entire design, make, test and analyse cycle of drug discovery, a breakthrough currently absent in industry.

We innovate across the entire workflow, including molecular optimisation using constrained reagent libraries, automated high-throughput synthesis of products, high-throughput analysis and purification, and more. Each of these issues are being addressed in the ChemTech Lab to build a fleet of autonomous platforms. These platforms have the potential to increase the throughput of molecular syntheses and testing by 250x per week when compared to conventional approaches as well as reducing time and labour costs of early-stage drug discovery by over 95%, resulting in cost savings of up to $14m per drug molecule entering clinical trials.

This research will enable faster and more effective lead optimisation in drug discovery, thereby enabling new targets that are currently economically unfeasible to research, such as rare disease and poverty-related disease. According to United Nations figures, more than 50% of the world's population growth between now and 2050 will be in just nine low-income countries, meaning that millions more people will suffer from these diseases with limited resources committed to addressing this problem. Furthermore, climate change-related disease migration will see many of these diseases spread further North and South, where many are ill-prepared to tackle these new diseases quickly. Through this research, these autonomous platforms are being developed and installed to greatly accelerate the discovery of new therapeutics, thereby increasing preparedness and futureproofing against new disease outbreaks and potential pandemics, dampening economic and societal effects.

Relevant projects: Autonomous Evolution of Molecular Fragments to Drug Candidates, Machine-Guided Molecular Growth for Drug Discovery.

Chemical Understanding, Modelling & optimisation

We develop and apply machine learning methods to accelerate the understanding and optimisation of chemical processes. Data-efficient approaches such as Bayesian optimisation and multi-task learning are used to explore high-dimensional reaction spaces, enabling rapid identification of optimal conditions while minimising experimental cost and material usage.

Beyond optimisation, we use machine learning for kinetic analysis and reaction simulation to extract mechanistic insight from experimental data. These models allow prediction of yield, selectivity and impurity formation, supporting the rational design of robust and scalable chemical processes.

This research addresses a key bottleneck in process development: the time and cost associated with empirical optimisation and late-stage failure. By maximising yield and selectivity while minimising impurity formation at source, our work reduces downstream purification requirements, shortens development timelines and lowers the overall cost of chemical and pharmaceutical manufacturing.

Relevant projects: A Sustainable Future Factory, ZeroShotAPI.

Reactor development and sustainable process development

We design and build autonomous reactor platforms for sustainable chemical process development, combining flow chemistry, laboratory automation and closed-loop optimisation. Using custom-designed and 3D-printed reactor systems, we generate high-density experimental data under precisely controlled conditions, enabling rapid process optimisation and robust model development.

Flow-based operation allows efficient exploration of reaction conditions while minimising solvent use, waste generation and energy consumption. Coupling these reactors with real-time analytics and machine learning-driven optimisation enables processes to be tuned for yield, selectivity and resource efficiency simultaneously, supporting sustainable and scalable chemical manufacture from the outset.

This work underpins our broader commitment to resilient and sustainable chemistry and directly informs the Enhancing Sustainability in Chemical Manufacture theme of the CDT in Resilient Chemistry, led by Connor. Together, these efforts aim to reduce the environmental footprint, cost and risk associated with chemical process development, helping to deliver manufacturing routes that are both economically and environmentally robust.

Relevant projects: ZeroShotAPI, ICEBERG, Enhancing Sustainability in Chemical Manufacture.