I've fine-tuned and evaluated AI model outputs across technical, creative, and business fields as an independent AI Training Specialist.
I build custom prompt templates and instruction sets for complex, multi-step LLM tasks, then score and correct outputs using RLHF principles to reduce hallucinations and improve factual accuracy. I also test edge cases and edge scenarios to uncover logical flaws and refine model performance.
Through contract and project work, I've designed domain-specific datasets for supervised fine-tuning and reviewed AI-generated text and visual prompts against strict formatting and style requirements.
I maintain prompt libraries that streamline content creation and repetitive workflows, drawing on machine learning concepts, advanced prompt design, and natural language processing.
