Dr David Mguni
/filters:format(webp)/prod01/channel_164/school-of-electronic-engineering-and-computer-science/media/eecs/professional_portrait.png)
Lecturer in Computer Science
Email: d.mguni@qmul.ac.ukRoom Number: Peter Landin, 4th floor
Profile
David Mguni is a Lecturer in Computer Science at Queen Mary University of London, where he is a member of the Centre for Multimodal AI and the Centre for Fundamentals of AI and Computational Theory. His research focuses on learning and decision-making under uncertainty, bringing together reinforcement learning, game theory, meta-learning, stochastic control and statistical learning. His work develops theoretical and practical approaches to intelligent systems that can learn, adapt and interact effectively in complex and uncertain environments, with particular emphasis on multi-agent systems and adaptive AI.
David has an established publication record at leading international conferences in artificial intelligence and machine learning, including ICML, ICLR, NeurIPS, AAAI and IJCAI. He serves as an Area Chair for ICML and ICLR, contributing to the peer-review process and scientific evaluation of research in the field. David holds a PhD in Applied Mathematics from University College London and previously studied at the University of Cambridge and the University of London. Before joining Queen Mary, he was a Principal Scientist at Huawei, where he led the Multi-Agent Systems research team in London.
Teaching
Distributed Systems — ECS656U / ECS796P
This module examines the principles and algorithms underpinning modern distributed computing, from cloud services and distributed databases to decentralised networks. Students explore how networked computers communicate, coordinate and maintain consistency despite communication delays, failures and incomplete information. Core topics include distributed processes, logical and vector clocks, synchronisation, distributed snapshots, replication, consensus and fault tolerance, alongside influential algorithms and protocols such as Lamport clocks, Chandy–Lamport snapshots, Paxos, Raft and Chord. The module combines theoretical foundations with practical system-design considerations, developing students' ability to analyse the behaviour, limitations and trade-offs of distributed systems.
Risk and Decision-Making for Data Science and AI — ECS7005P
This module develops the mathematical and computational foundations of reasoning and decision-making under uncertainty. Topics include probability, Bayesian inference, quantitative risk, causal reasoning, Bayesian networks, decision theory and expected utility. Students investigate how evidence informs beliefs, how uncertainty influences decisions, and why prediction and correlation alone may be insufficient for effective decision-making. The module combines theoretical reasoning with practical probabilistic modelling, including the construction and analysis of Bayesian networks
and decision models using tools such as PyAgrum. Applications span healthcare, engineering, finance and artificial intelligence, equipping students to critically evaluate risk and make informed decisions under uncertain or incomplete information.
Research
Research Interests:
David's research focuses on learning and decision-making under uncertainty, combining reinforcement learning, game theory and multi-agent learning to understand how intelligent systems learn, interact and adapt in complex environments. His work investigates cooperation, competition, equilibrium behaviour, exploration and reward design, alongside the development of safe, robust and efficient learning algorithms. He is particularly interested in the mathematical foundations of learning and how the structure of games, optimisation problems and dynamical systems can inform the design of more effective intelligent systems.
A second major strand of his research concerns meta-learning, adaptive intelligence and self-improving AI, investigating how systems can transfer knowledge across tasks, adapt from limited experience and improve their own learning processes. His interests include few-shot and continual learning, probabilistic modelling, stochastic control and stochastic differential equations, as well as reasoning and reliability in large language models and agentic AI. Across these areas, his overarching aim is to develop principled methods that enable intelligent systems to learn and make decisions reliably in uncertain and changing environments.
Supervision
I supervise research projects broadly concerned with machine learning, artificial intelligence and decision-making under uncertainty. I am most interested in supervising projects that pose a clear scientific question about how or why an intelligent system behaves as it does.
Areas of particular interest include:
- Reinforcement learning and multi-agent systems: cooperation and competition, equilibrium learning, safe and robust reinforcement learning, reward design, exploration and coordination between heterogeneous agents.
- Meta-learning and self-improving AI: learning how to learn, adaptive learning, few-shot learning, continual improvement and autonomous generation or selection of useful learning experiences.
- LLM reasoning and agentic systems: mathematical reasoning and theorem proving, strategy and plan evaluation, uncertainty and abstention, and communication between multiple LLM-based agents.
- Games, optimisation and learning algorithms: the behavioural and geometric structure of games and optimisation algorithms, algorithm selection and efficient equilibrium computation.
Projects can range from focused empirical investigations using existing models and datasets to more technically ambitious work involving new learning algorithms, mathematical analysis or theoretical results. Projects combining theoretical and empirical approaches are also welcomed.