Making AI Speak Your Research Language: An Introduction to Fine-Tuning LLMs
A three-hour, in-person, hands-on workshop where participants learn how to fine-tune a large language model (LLM), using Python, for their own research tasks.
On this page:
Large Language Models (LLMs) are AI computer programs trained on massive amounts of data to understand, summarise, translate, and generate human-like language. They offer broad capabilities but often lack the nuanced understanding required for specialised research tasks. By fine-tuning LLMs on custom datasets, we can alter their output and dramatically improve their performance on domain-specific tasks. This workshop provides a step-by-step framework for using Python to adapt LLMs to your specific domain context, from data preparation all the way through to model validation.
Participants need some familiarity with Python scripting (working with dictionaries, lists, functions, and loops). See Programming with Python for an introductory Python workshop.
Event details
Dates for 2026:
- Wednesday 18 November, 1–4pm, in-person at the City Campus.
Learning outcomes
Participants who attend this workshop will be able to:
- Explain how fine-tuning, prompting, and retrieval-augmented generation differ, and which is most appropriate for a given research task.
- Prepare and format a custom dataset for fine-tuning.
- Fine-tune an LLM using a parameter-efficient method such as LoRA.
- Select and apply appropriate methods to evaluate a fine-tuned model's performance.
After the workshop
- Researchers and postdoctoral students are encouraged to engage with their peers by joining Hacky Hour to post questions and exchange ideas.
Contact
Research Data Support Services
Email: researchdata@auckland.ac.nz