Prompt Engineer / AI model trainer (RAG and LLM prompt engineering)
Text
Prompt + Response Writing (SFT)Worked on designing, engineering, and optimizing prompts to enable large AI models to perform accurate attribution analysis and generate coherent responses. Applied RAG (Retrieval-Augmented Generation) technology to improve the quality and relevance of generated answers in specialized domains like power grid operations and knowledge management. Leveraged Qwen-72B and BGE-M3 models for natural language understanding, function calling, and workflow orchestration in production systems. • Developed and tested prompts for question answering and attribution tasks leveraging domain-specific data. • Conducted evaluation and tuning of model responses to ensure relevance and accuracy in specialized technical workflows. • Iteratively improved labeling workflows using user Q&A logs and feedback for enhanced model performance. • Collaborated to build and maintain internal tools for prompt engineering and real-time model evaluation.