Research team
Diotima was built by educators, AI researchers and governance specialists working together and does not treat compliance as an afterthought.
Diotima began with a workload problem its founder lived through.
Siobhan Ryan retrained as a teacher and spent her evenings and weekends marking work and writing feedback. She knew feedback was the part that actually moved her learners forward. She also knew there was no version of her week in which she could give all of them enough of it.
That problem is not confined to schools. It is sharper in professional education, where one trainer can be responsible for hundreds of learners across a programme and feedback often stops at pass or fail. The learner is told whether they got through. They are not told what to do differently.
Siobhan could see that AI could close that gap, but only if it was built to a standard that training and education organisations could actually adopt:
Safe, transparent and compliant with the rules coming into force across Europe.
Diotima is the result.
Diotima brings together three groups:
The School of Education at Trinity College Dublin grounds the product in what is actually known about how people learn.
The School of Computer Science defined how we build and test AI.
The learning technologists and design specialists in the Learnovate Centre shape the user experience.
Through Learnovate, we work directly with practising educators, testing what we build with the people who will use it rather than assuming we already know.
Diotima is built to meet the EU AI Act, GDPR, the EU Digital Services Act and the European Accessibility Act.
Because a system that assesses learners is treated as high risk under the EU AI Act, that obligation does not sit with us alone.
It reaches the organisations that deploy it.
Building for that from the start is what allows training and education organisations to adopt Diotima without taking on a compliance problem on its own.
Co-founder of Digitary, previously Head of Education for TCS in Europe, and chair for 2026 of the Responsible AI for Learning (RAIL) community of practice.
A former teacher who lived the feedback problem Diotima was built to solve.
Keeps the product grounded in evidence about how people actually learn.
Principal Investigator and
Computational Linguist
Diotima Chief Technical Advisor