Human-in-the-Loop Decision-Making for Lipid Nanoparticle Reactor Design

Multi-fidelity Bayesian optimisation (MFBO) navigates this cost–accuracy trade-off, but treats all fidelity sources as computational. In practice, pharmaceutical experts hold knowledge that simulations limitedly encode, such as manufacturability constraints, regulatory preferences, and intuitions about feasibility and operating windows.

Based on our previous work combining computational fluids dynamics (CFD) with multi-fidelity Bayesian optimisation for pharmaceutical reactor design (Savage, Basha et al., Nature Chemical Engineering, 2024) and geometry-conditioned surrogate modelling for multiphase flows (Ramlau, Basha, Matar et al., ICML AI4Physics workshop, 2026), this project will develop a human-in-the-loop multi-fidelity framework that treats the process expert as a query-able fidelity source within the optimisation loop.

Aims of this project are: i) formulate an adaptive agent policy that reasons jointly over surrogate queries, complex CFD simulations, human expert elicitation, and generative design proposals, learning when each action yields the highest information gain relative to its cost; ii) develop preference-based elicitation, such as pairwise comparisons, feasibility judgements, soft constraint injection, capturing expert knowledge; iii) design interpretable uncertainty communication that surfaces why the model is uncertain (geometric novelty, surrogate- CFD disagreement, extrapolation), enabling informed trust-or-override decisions; iv) validate on CIJ mixer geometry optimisation for LNP nanoprecipitation, benchmarking against standard MFBO.

Multi-output Gaussian processes will model correlations across fidelity levels and output quantities (particle size moments, mixing efficiency), providing calibrated uncertainty for the acquisition strategy. Preferential GP models will encode human feedback through pairwise comparison likelihoods, enabling qualitative expert judgements to update the posterior without scalar scores. The agent's action-selection policy will be formulated as a contextual bandit trained to maximise information gain per unit cost. Conditional flow matching models will propose candidate reactor geometries conditioned on target specifications, with learned feasibility classifiers filtering proposals that would fail to mesh or converge before they consume simulation budget.

Every new biological therapy must be manufactured, and manufacturing design remains one of the slowest steps in bringing treatments to patients. The core obstacle is not a lack of models or a lack of expertise, it is that models and experts currently operate in sequence rather than in collaboration. This project changes that: an agent that learns when to compute, when to generate, and when to ask the expert, making human judgement a resource the algorithm actively seeks rather than a checkpoint it passively waits for. The immediate impact is faster, cheaper development of LNP-based therapies; the lasting contribution is a general framework for human-AI collaborative design of any engineered biological system.

Desirable Student Background

Good foundations in probabilistic methods, Gaussian Processes or Bayesian optimisation. Knowledge of Partial differential equations or physics-based models will be useful. Proficiency in Python and mathematical maturity (linear algebra, probability, statistics) is needed. Ideally interested in collaborative work with both computer scientists and domain experts.

How to apply

Please apply under the University of Manchester application portal and select 'PhD in Artificial Intelligence' (OAA Application form).

When applying, you’ll need to specify the full name of this project, the name of your supervisor, if you already having funding or if you wish to be considered for available funding through the university, details of your previous study, and names and contact details of two referees.

Your application will not be processed without all of the required documents submitted at the time of application, and we cannot accept responsibility for late or missed deadlines. Incomplete applications will not be considered.

You will be asked to upload the following supporting documents:

• Final Transcript and certificates of all awarded university level qualifications
• Interim Transcript of any university level qualifications in progress
• CV
• Supporting statement: A one or two page statement outlining your motivation to pursue postgraduate research and why you want to undertake postgraduate research at Manchester, any relevant research or work experience, the key findings of your previous research experience, and techniques and skills you’ve developed. (This is mandatory for all applicants and the application will be put on hold without it).
• Contact details for two referees (Please ensure referee email addresses are official university or workplace accounts, as we may need to verify the references.)
• English Language Certificate (if applicable)

We actively encourage applicants from diverse career paths and backgrounds and from all sections of the community, regardless of age, disability, ethnicity, gender, gender expression, sexual orientation and transgender status.

We also support applications from those returning from a career break or other roles. We are dedicated to supporting work-life balance and offer flexible working arrangements to accommodate individual needs. Our selection process is free from bias, and we are committed to ensuring fair and equal opportunities for all applicants.

We are dedicated to supporting work-life balance and offer flexible working arrangements to accommodate individual needs, including a part-time study option.

Join us and help grow the UK’s national capability in artificial intelligence, part of the UK’s modern industrial strategy. Everyone is welcome if you are passionate about uncertainty quantification, decision making with humans in the loop and decision making for machine learning systems and their applications in Science and Engineering, and looking to develop your research and innovation skills at the doctoral level to further your career in the AI sector.

Equality, diversity and inclusion is fundamental to the success of The University of Manchester, and is at the heart of all of our activities. We know that diversity strengthens our research community, leading to enhanced research creativity, productivity and quality, and societal and economic impact. (Equality, diversity and inclusion | The University of Manchester)

We actively encourage applicants from diverse career paths and backgrounds and from all sections of the community, regardless of age, disability, ethnicity, gender, gender expression, sexual orientation and transgender status.

Funding Notes

Fully funded AI UKRI CDT 4 year program; The Project will be based in University of Manchester.

The UKRI AI CDT in Decision Making for Complex Systems is part of the TechFirst offering stipends of £30,000 for students eligible for home fee status. The start date is September 2027.

References

Machine learning-assisted discovery of flow reactor designs: https://www.nature.com/articles/s44286-024-00099-1
ICML AI4Physics 2026: https://arxiv.org/abs/2606.16587