AI Trainer/Web Search Evaluator (Independent Contractor)
I contributed to the training of AI models for multiple large language models (LLMs) by evaluating outputs, formulating prompts, and assessing responses. This work involved both generating novel scenarios for the models and conducting deep factual accuracy checks regarding scientific and non-scientific topics. The role required significant research, including referencing peer-reviewed papers and examining models' reasoning and reliability. • Assessed AI model responses for accuracy and alignment with given prompts. • Designed prompts and scenarios to challenge model limitations and boundaries. • Conducted evaluations, including Red Teaming tasks, to detect undesirable behaviors. • Verified scientific content by cross-referencing with source publications.