Walk me through the process of fine-tuning a pre-trained Large Language Model (LLM) for a domain-specific text classification task at Thomson Reuters.