AI/ML Data Trainer | Prolific & Amazon Mechanical Turk
I trained large language models (LLMs) by performing A/B testing, evaluating generative AI outputs, and annotating content for safety and alignment. I provided detailed ranking and feedback on text and image outputs to improve overall AI quality. I also labeled UI elements and analyzed various content domains for accuracy and safety. • Focused on creative writing, red teaming, safety protocols, news analysis, and booking workflow interactions. • Contributed to identifying and reducing harmful outputs in LLMs through evaluation and annotation. • Provided feedback for UI elements and agent performance across browsers and operating systems. • Used annotation interfaces provided by Prolific and Amazon Mechanical Turk platforms.