Skill

LLM Fine-Tuning

Adapting open-weight language models through fine-tuning on custom-built datasets, including via distillation.

I build fine-tuning pipelines for open-weight language models, including fine-tuning Gemma 3 on a custom dataset built through distillation from a frontier LLM. The flagship project here is Manzoni AI, a model that writes in the style of Alessandro Manzoni.

Related projects

2025Completed

Manzoni AI

A LoRA fine-tuning of Gemma 3 4B IT that translates modern Italian into the literary style of Alessandro Manzoni, built on a custom parallel dataset distilled from I Promessi Sposi.

PythonLLM Fine-Tuning

Related articles

Jun 4, 20258 min read

Fine-Tuning a Language Model to Write Like Manzoni

How I built a custom parallel dataset through distillation and adapted Gemma 3 4B to perform literary style transfer in nineteenth-century Italian prose.

#llm#fine-tuning#nlp#italian