LLM Sentiment Data Labeler (FinSentiment Project)
As part of the FinSentiment: LLM-driven Financial Sentiment Dashboard project, I automated sentiment scoring using a fine-tuned large language model (LLM) on finance datasets. I built a Python/PyTorch dashboard that analyzes over 500 daily finance articles, making use of labeled sentiment data to enhance model accuracy. My work significantly reduced manual analysis time and supported better decision-making for traders. • Developed and implemented data labeling protocols for finance article sentiment analysis. • Performed fine-tuning of an LLM using labeled datasets to optimize sentiment predictions. • Evaluated model outputs and provided feedback to improve classification reliability. • Enabled automation workflows by curating and preparing training data for NLP tasks.