We fine-tuned the LLaMa 3.1 8B model to create a sarcastic chatbot, structuring a dataset with three key columns: System Prompt, User Input, and Assistant Response, and using MonsterAPI for deployment. The fine-tuning process involved data preprocessing, adjusting training parameters, and training the model to recognize sarcasm based on user input and system prompts. After deployment, users tested the bot and provided feedback, helping us tweak the responses in the dataset for a more balanced experience. The chatbot was successfully deployed as an API endpoint, allowing users to interact with it in real-time, and the process showcased the feasibility of creating unique tone-based chatbots using available tools and resources.