BuddhAi Project - A multimodal Socratic AI tutor for novice programmers
Programming and algorithmic knowledge are foundational skills for students in technical fields, fostering problem-solving, computational thinking, and systematic reasoning in complex and dynamic environments. However, novice programmers often struggle with autonomy and decision-making, hindering the development of learner agency and self-efficacy. This challenge can lead to ‘blank canvas paralysis’ as students grapple with setting up environments, mastering tools, and debugging. Students increasingly turn to general-purpose LLMs like ChatGPT or Copilot. While these models provide fast, answer-driven support, they lack pedagogical grounding and offer limited scaffolding for thought and reflection. Meanwhile, teachers gain little insight into students' learning processes, limiting their ability to provide targeted support. At Aalborg University (AAU), we believe addressing the problem can only be achieved as an interdisciplinary, collaborative effort between technical and educational staff working across three departments and integrating a programming LLM tutor into two relevant case educational programs spanning two faculties before further scale-up. The project delivers the final tool, a pedagogical framework with new course materials and exercises, and an AAU policy document will allow continuous use and scale-up. The project will be evaluated on two different cohorts Medialogy (Technical Faculty of IT and Design ) and Clinical Science and Tech (Faculty of Medicine). Funded by NOVO Nordic Foundation PREPARE programme
BuddhAi Project - A multimodal Socratic AI tutor for novice programmers
Programming and algorithmic knowledge are foundational skills for students in technical fields, fostering problem-solving, computational thinking, and systematic reasoning in complex and dynamic environments. However, novice programmers often struggle with autonomy and decision-making, hindering the development of learner agency and self-efficacy. This challenge can lead to ‘blank canvas paralysis’ as students grapple with setting up environments, mastering tools, and debugging. Students increasingly turn to general-purpose LLMs like ChatGPT or Copilot. While these models provide fast, answer-driven support, they lack pedagogical grounding and offer limited scaffolding for thought and reflection. Meanwhile, teachers gain little insight into students' learning processes, limiting their ability to provide targeted support. At Aalborg University (AAU), we believe addressing the problem can only be achieved as an interdisciplinary, collaborative effort between technical and educational staff working across three departments and integrating a programming LLM tutor into two relevant case educational programs spanning two faculties before further scale-up. The project delivers the final tool, a pedagogical framework with new course materials and exercises, and an AAU policy document will allow continuous use and scale-up. The project will be evaluated on two different cohorts Medialogy (Technical Faculty of IT and Design ) and Clinical Science and Tech (Faculty of Medicine). Funded by NOVO Nordic Foundation PREPARE programme