About Us
L2-Bench is an inter-disciplinary, collaborative project, initiated at Oxford University Press in 2025, and launching in 2026. Our team actively maintains the L2-Bench leaderboard and open benchmark dataset, with future releases planned to help support an AI for Education evaluation ecosystem.
Team
Meet the core team of pedagogy experts, data scientists and AI researchers that developed the first release of L2-Bench:
Lead Data Scientist
James Edgell is a Lead Data Scientist at Oxford University Press. He is responsible for the design and build of machine learning systems that serve learners and teachers of English Language, and is leading research into AI benchmarking methodologies that evaluate AI's effectiveness in language education. James studied Physics at the University of Warwick and postgraduate Geophysics at the University of Leeds. He began his data science career building ML systems for personalisation at the UK's largest hotel chain, before leading analytics teams in North America's airlines industry, building expertise in Natural Language Processing. James then joined OUP to improve learning outcomes, prioritising quality and safety in the development of trustworthy AI.
AI Incubation Manager
Isaac is an AI Incubation Manager at Oxford University Press, specializing in designing and deploying AI solution architectures. With expertise spanning information architecture, machine learning, and natural language processing, he architects foundation model adaptations and agentic AI systems. Isaac holds an MLIS from the University of Washington and brings extensive experience from Microsoft's Cloud + AI division. He focuses on transforming complex AI capabilities into intuitive, production-ready systems that bridge data science and business applications. Previously a Language Information Architect optimizing datasets for model training, Isaac combines technical knowledge with strategic implementation to drive innovation in enterprise AI deployment.
Head of Pedagogy Research
Ben is Head of Pedagogy Research at Oxford University Press. He is responsible for ensuring a clear research-informed pedagogical approach underpinning Oxford ELT courses and learning materials. His current priority is developing a pedagogical-based approach to using emerging technology in order to help teachers and learners in their language learning. He studied Linguistics at the University of York and Applied Linguistics at the University of Edinburgh. He has been a teacher, trainer and lecturer, in schools, universities and the British Council, in various countries around the world, in Asia, Africa and Europe.
Learning Designer and Pedagogy Specialist
Danielle is a Learning Designer and Pedagogy Specialist at Oxford University Press, where she blends pedagogical insights with practical learning design to create learner-centred language learning experiences. Her work draws on the science of learning and design thinking to ensure solutions are both research-informed and effective. She holds a Cambridge DELTA and brings a broad background in educational publishing, having held roles at FlashAcademy, Astrid Education, Springer Nature, and EF Education First.
AI Evaluations Researcher
Wm. Matthew Kennedy is a Marie SkÅ‚odowska-Curie Postdoctoral Fellow at the Oxford Internet Institute. He researches AI ethics, sociotechnical safety, and social impacts, focusing on three areas in particular: AI evaluation and red-teaming methodologies; knowledge production; and decolonial AI. Before joining the OII, he received his PhD from the University of Sydney in the history of colonialism, international law, and “scientific” governance. After spending time in the technology industry, he became involved in AI evaluations, eventually leading a team tasked with structuring evaluations of AI systems, collaborating with external colleagues to produce research, and advising public officials on AI policy.
Supporting Team
- Megan Gericke, AI Incubation Project Manager, Oxford University Press
- Martin Ku, AI Context Engineer, Oxford University Press
- Dorian McCree, Director Learner Progress & Analytics, Oxford University Press
Collaborators and Contributors
We are grateful to the following collaborators and contributors who have been part of the L2-Bench journey:
Research Collaborators
Professor Elizabeth Wonnacott, Department of Education, University of Oxford
Dataset Collaborators
Beatrice Segura Harvey, ELT specialist
Contributors
We would like to thank all 39 postgraduate participants and organisers of the 2025 University of Birmingham “PGT SHAPE AI Challenge” for their hard work and valuable contributions in helping us iterate on early versions of L2-Bench. Participants from winning and runner-up teams who consented to be named are included in methods paper acknowledgements; their inclusion does not imply endorsement of the research.
We would like to thank all 221 education practitioners who contributed their time and expertise to the 2026 “OUP Global Practitioner Challenge” study, helping us to understand the validity of L2-Bench prior to the first open release, and for helping us to identify where to focus our efforts in future iterations. Practitioners who qualified and consented to be named will be included in forthcoming results paper acknowledgements; their inclusion does not imply endorsement of the research.
Connect with Us
We welcome any general enquiries, support opportunities, and feedback about the L2-Bench project. Whether you're a practitioner interested in using our benchmark, an institution considering AI tools for language education, or a researcher looking to build on this work, we'd love to hear from you. Please get in touch via the ”Register Interest” webform below.
Research Collaboration and Benchmark Support
L2-Bench is a collaborative research project. To help us in supporting an AI for Education evaluation ecosystem, you can ”Register Interest” in co-developing or funding the next phase of this work, or advocate L2-Bench to help spread the word.