LLM Workflow Prompt Engineering and Data Annotation
In my AI-Powered Customer Insight & Feedback System project, I designed and refined workflows involving prompt engineering and feedback classification using large language models (LLMs). The core work involved fine-tuning and validating LLM outputs to generate concise summaries from unstructured feedback data. I labeled and evaluated prompts and responses to enhance LLM performance and ensure actionable customer insights. • Conducted prompt engineering and response rating for feedback summarization tasks. • Labeled feedback categories (e.g., product quality, service issues) in community feedback data. • Employed sentiment analysis and NER for accurate text annotation. • Iteratively tested and improved model prompts via A/B validation cycles.