QoE-Driven Agentic AI for Automated Bug Discovery in Video Games (Collaborative Doctorate with Sony Interactive Entertainment) University of Surrey in United Kingdom

This collaborative PhD project between the University of Surrey (School of Arts, Humanities and Creative Industries) and Sony Interactive Entertainment (SIE) focuses on developing a Quality of Experience (QoE)-driven agentic AI framework for automated bug discovery in video games. Video games are increasingly complex and updated frequently, making manual quality assurance (QA) challenging and often insufficient for covering vast interactive state spaces. Automated approaches typically detect anomalies but lack clarity in issue explanation and prioritisation based on player impact.

The research will implement a multi-agent workflow with three distinct roles: the Explorer agent actively probes games to discover failures; the Inspector agent verifies and reproduces candidate issues, capturing minimal but sufficient evidence (inputs, states, clips, logs); and the Reporter agent produces structured bug reports suitable for triage. A key innovation is the explicit QoE severity model, which learns to rank bugs based on predicted player impact—such as frustration, immersion disruption, fairness issues, comfort, or usability—using lightweight human feedback and QA expertise.

Evaluation will focus on improvements over scripted and random baselines, using metrics like confirmed unique bugs per hour, reproducibility rate, evidence quality, and correlation between severity ranking and human judgements. The student will benefit from SIE co-supervision, industry-informed glitch taxonomies, and reporting requirements, all under appropriate data, IP, and publication governance. The project aligns with Surrey’s GAIN programme and games provision within SAHCI, supporting standards-oriented impact through the supervisory team’s engagement with ITU-T work on gaming QoE.

Funding is fully and directly provided for this project, covering Home or International fees for 42 months. The UKRI standard stipend is £21,805 per annum for the academic year 2026/27, with an RTSG of £1,500 per year and potential additional conference funding up to £3,000, subject to approval.

Applicants must meet the minimum entry requirements for the PhD programme at University of Surrey. Preferred backgrounds include Computer Science, Games Technology, Digital Media, or related disciplines. Strong programming skills (preferably Python) and experience in machine learning/AI or software engineering for interactive systems are required. Experience in game development or design using Unity, Unreal, or Godot is desirable. Up-to-date knowledge of generative AI, vision-language models, and agent-based AI systems is advantageous. Candidates should be able to design and conduct human-centred evaluation experiments, including QoE, usability, and preference studies, and demonstrate excellent communication skills and the ability to collaborate with industry partners under confidentiality constraints.

The position is open to UK and international candidates, with a start date in October 2026. To apply, discuss your project with a prospective supervisor and submit your application via the Innovative Media Technology PhD programme page. In place of a research proposal, upload a document stating the project title and the name of the relevant supervisor.