Contributed to the Open-Finance-Lab/FinLLM-Leaderboard repository by developing comprehensive tutorial documentation for a zero-shot approach, focusing on FLARE-FIQASA datasets and the Llama-3.2-1B model. The work detailed dataset usage and sentiment classification tasks, incorporating performance logging to support analysis and reproducibility. The documentation addressed partial matching evaluation, concurrency considerations, and token streaming to enhance workflow efficiency. By emphasizing clarity and onboarding, the contribution enabled engineers and contributors to quickly understand and optimize zero-shot machine learning workflows. This effort leveraged skills in documentation, machine learning, and natural language processing, with all content authored in text-based formats for accessibility.