AC

Explore

06 / 07

Outside of work

Blogs

Technical notes and reflections on software, systems, artificial intelligence, and the process of learning.

Diagram showing LoRA adapter training and the merged model weights after training

Large language models

Apr 4, 2025

LoRA, an efficient approach to fine-tuning a Large Language Model

A practical explanation of why parameter-efficient fine-tuning matters, how Low-Rank Adaptation updates a small set of trainable weights, and how I applied it while building the MedLam medical question-answering model.