A configuration-based command-line tool for launching LLM fine-tuning experiments on custom data. It orchestrates prompts, open-source LLMs, and optimization strategies through YAML files to control the experimentation pipeline end-to-end. Supports popular model families including Llama2, Mistral-7B, FLAN-T5, Falcon, and Zephyr, with techniques such as LoRA and QLoRA for parameter-efficient fine-tuning, along with ablation studies and unit testing workflows.