Join us online or in-person at University of Glasgow on at 2pm on Monday, 15 December 2025 for our final Education Seminar of 2025 where Eyman Alyahyan will present: Pedagogical Prompt Engineering Protocol Applied to LLM Feedback in Introductory Programming.
Abstract:
Many researchers and educators are exploring how Large Language Models (LLMs) could be used to support feedback practices in computing education. Yet the prompts that guide these systems are often ad-hoc and not grounded in pedagogy. What if we could design prompts that enable LLMs to produce feedback that supports student learning?
In this seminar, Eyman Alyahyan introduces the Pedagogical Prompt Engineering Protocol (PPEP)—a systematic, theory-informed methodology for developing effective educational prompts. PPEP served as the methodological foundation for this work. Applying the protocol led to the PPE-LLM framework, a set of 10 components that translate the protocol into a clear design structure for building prompts. These components give teachers a ready-to-use scaffold for creating a well-structured prompt. That prompt can then be used to generate formative feedback using LLMs that supports student learning in introductory programming.
The talk shows how moving from principles to practice can make LLM-generated feedback more consistent and pedagogically meaningful. It also offers a transparent, repeatable approach for researchers and developers to create future prompt frameworks and LLM-based feedback tools grounded in pedagogy rather than intuition.