Mini-project
Insurance
Fine-tuned claim note summariser
Fine-tune a small model to summarise surveyor notes, and compare it with a large model.
The brief
Claims managers read long surveyor notes to decide each claim. Fine-tune a small open model to write a short, structured summary of a surveyor's note, and compare it with a large hosted model using an evaluation set, on quality, speed and cost. Good looks like: a training set of note and summary pairs, a clear evaluation rubric, and a recommendation.
What you will build
- Create note and summary pairs from claim records
- Fine-tune a small model with LoRA
- Generate summaries with both models
- Score with a rubric, an LLM judge and spot checks
- Compare quality, speed and cost
Free, open data to build it with
Check each dataset's licence before you publish your work. Your mentor gives your team its own version of the brief (data, scope or region), so every team's project is its own.