Postdoctoral Research Associate - Strand, London, WC2R 2LS

Postdoctoral Research Associate - Strand, London, WC2R 2LS

About us

The Department of Informatics is looking to recruit a Postdoctoral Research Associate (PDRA) to join the lab of Dr. David Watson in the Department of Informatics, based at Bush House in central London.

The successful applicant will work on PRECISE-AI (Probabilistic Reasoning with Circuits for Safe and Explainable Artificial Intelligence), a research project funded by the Engineering and Physical Sciences Research Council. King’s Department of Informatics is a large, interdisciplinary department with a vibrant research culture and long history of cutting-edge AI research.

The postholder will benefit from access to King’s CREATE high-performance computing cluster, as well as mentoring and professional development opportunities provided by the King’s Centre for Research Staff Development.

About the role

The PDRA will carry out research on probabilistic circuits (PCs) and tree-based machine learning methods for generative modelling, with a particular focus on responsible AI. The project builds on adversarial random forests (ARFs), an unsupervised learning framework developed by the PI that combines classification and regression trees with adversarial training to produce fast, flexible, and interpretable generative models that can be compiled into PCs for tractable probabilistic inference.

PRECISE-AI has three core objectives: (1) designing differentially private PC architectures that protect against membership inference attacks; (2) developing novel explainability methods that characterise model outputs in terms of necessary and sufficient conditions; and (3) building circuit-based auditing tools for causal fairness criteria. These pillars of responsible AI have been extensively studied in supervised learning but remain largely unexplored in the generative setting. Theoretical aims of the project include deriving statistical guarantees (e.g., minimax convergence) and computational tractability results (e.g., polytime complexity). PRECISE-AI will focus in the first instance on biomedical applications, using genomic and clinical data from large cohort studies on immune-mediated inflammatory diseases including psoriasis and rheumatoid arthritis.

The PDRA will lead on research across all three objectives, taking first authorship on associated publications and playing a central role in software development and dissemination. Target publication venues include NeurIPS, ICML, AAAI, and the Journal of Machine Learning Research. The postholder will collaborate with industry partners at Meta and academic collaborators at UCL and Cambridge. The PDRA will also have the opportunity to organise workshops at leading international conferences. The project will produce open-source software packages for Python and/or R, as well as user-friendly web applications aimed at practitioners across healthcare, finance, and industrial engineering.

This is a full- time post (35 hours/week), and you will be offered an a fixed term contract for 24 months ending no later than 30th June 2028.

Research staff at King’s are entitled to at least 10 days per year (pro-rata) for professional development. This entitlement, from the Concordat to Support the Career Development of Researchers , applies to Postdocs, Research Assistants, Research and Teaching Technicians, Teaching Fellows and AEP equivalent up to and including grade 7. Visit the Centre for Research Staff Development for more information.