We conduct empirical studies of software engineering to provide credible evidence for decision making in software development, maintenance and deployment. We offer expertise in code analysis, participatory action research, multi-vocal literature reviews and case studies. 

We study automated software engineering for fog computing, continuous integration of service-oriented systems, and web API/mashup development and maintenance.

From digital health devices, social network platforms, to smart living infrastructures, software systems are primarily designed and built to address humans needs and ambitions. A wide variety of human-centric issues impact the design, development, and use of software systems, including the demographics (age, gender, education, incomes), abilities (physical and mental), and values (privacy, security, environmental care, equity, and autonomy) of software practitioners, end-users, and wider stakeholders. Our work centres on the investigation of how human-centric issues affect SE practice and we develop methods, tools, and techniques to explicitly consider and address these issues through the SE decision making process, from systems requirements analysis, to design, implementation, testing, validation, deployment, maintenance, retirement, and disposal. 

We study legacy systems to better understand their complexity, and so better manage the maintenance and evolution of such legacy systems, and also reduce or minimise technical debt. We study such systems through analysis of multiple information sources. The ultimate aim is to de-risk such systems through supported transformation to modern architectures and techniques. 

Requirements changes during software development create risk to cost, schedule and quality but also represent an opportunity to add value. How can we predict and manage requirements change? We have looked at predicting requirements change from project and repository data and making use of expert judgement. 

How can we use machine-based search techniques to solve different optimisation problems from various software engineering areas? Faced with a huge search spaces and the difficulty of using a human-based search, often we can find the solution(s) in metaheuristic searches and evolutionary algorithms. We have used genetic algorithms, simulated annealing, taboo search and other techniques to solve problems in software release planning and in automated software refactoring. 

We develop, evaluate and deploy technologies and practices for safe, fair, and trustworthy AI. 

For more than a decade, we have performed industry-relevant, applied research on software testing and quality assurance. Most our projects are in joint collaboration with industry and with various academic/industrial collaborators in the UK, Europe and elsewhere in the world. All our projects result in technology transfer, scientific/industrial papers, and completion of several MSc / PhD theses. 

We study the human and technical aspects of team-based software innovation, including remote work. The includes innovation practices (e.g., design thinking and agile), ‘thinking styles’, the tools for online collaboration, and team dynamics.