Jake Renzella | Senior Lecturer
University of New South Wales

Dr Jake Renzella is a Senior Lecturer in Computer Science, and Co-Head of the Computing and Education research group in the School of Computer Science Engineering at the University of New South Wales. Jake's research is in the field of pedagogically-sound generative AI for education. His work has been published in premier conferences and journals such as the SIGCSE TS, and the International Conference on Software Engineering. Importantly, Jake's work is deployed in the field, via the Debugging C Compiler and it's generative AI debugger, used over 500,000 times to support novice programmers, and Formatif used at several Australian and New Zealand universities with over 260,000 students. Jake's recognitions include:

The 2024 Australian Financial Review Higher Education Innovation Award
Various UNSW Teaching awards
Associate Fellow of the Higher Education Academy
Early Career Academic member of the Australasian Association for Engineering Education

Previously, Dr Renzella was a Research Fellow at the Applied Artificial Intelligence Institute (A²I²).

Appearances:

Day 2 @ 15:30

Generative AI that gets it: Building pedagogically-sound AI

Large Language Models (LLMs) such as ChatGPT are designed to be helpful assistants, and their capabilities are rapidly improving
There is huge potential for using LLMs in education, but current models often help too much and don’t align with effective teaching practices.
Just as we have shaped current LLMs to be helpful, we can shape LLMs to behave as effective educational agents.
We will discuss what properties make a "good" pedagogical LLM, such as: Self-efficacy facilitation, reducing over-helpfulness, and Socratic Guidance.
Participants will see a live demo of a real pedagogical LLM (used over 500,000 times) and compare its behaviour to standard models like ChatGPT. These models can also run locally on institutional premises, addressing privacy concerns.

Jake Renzella, Senior Lecturer, University of New South Wales

last published: 29/Apr/25 22:35 GMT

back to speakers